Launched in 2005, today YouTube is one of the most popular social media platforms globally, with over 2 billion monthly active users. YouTube is the perfect platform for all kinds of video content, whether it’s related to education, guides, entertainment, news reports, etc. Over the years, YouTube has seen continuous growth in its user base from people of different genders, demographics, geographical reach, and more.
In this article, we are going to explore YouTube’s User statistics, including its monthly active users, audience size by country, gender distribution, and more.
Top YouTube User Stats
In 2024, YouTube has 2.9 billion monthly active users.
India is the leading country with the highest number of YouTube users, 476 million.
54.3% of YouTube users are Male, while 45.7% of the users are Female.
The average time spent by users in the US accessing YouTube is 48.7 minutes.
Reports suggest that YouTube’s monthly active users are projected to reach 3.1 billion in 2025.
Users aged 25 to 34 years old have the highest share of YouTube users at 21.3%.
In South Korea, the monthly time spent on the YouTube mobile app per user is 45 hours and 49 minutes.
YouTube Monthly Active Users
YouTube is the ultimate destination for individuals looking for entertainment, learning, and finding tutorials on various topics, advertisements, etc. This results in YouTube attracting a large number of users every month. In 2024, the platform reached a total of 2.9 billion monthly users. In Fact, reports also suggest YouTube will cross a milestone of 3 billion monthly active users in 2025.
Below, we have mentioned a table showcasing the growth in YouTube’s monthly active users from 2010 to 2025:
Individuals that fall under the age category of 25 to 34 are most likely to access YouTube with a total 21.3% share of users utilizing the platform. The Age group 35 to 44 is the second leading age category with 17.5% of users, while 18 to 24 years old is the third leading age group with 15.5% of YouTube users. Overall, young adults and early middle age users are more likely to access YouTube compared to other age groups. There is a significant drop in the share of users, after the age of 55.
Below we have mentioned a table showcasing YouTube users by Age Group:
Age Group
Share of Users
18 to 24
15.5%
25 to 34
21.3%
35 to 44
17.5%
45 to 54
12.5%
55 to 64
9.2%
Over 65
9.2%
YouTube Users worldwide by Age group and Gender
As of April 2024, 12.1% of the global YouTube audience was composed of Male users aged between 25 to 34 years old, along with 9.4% of female users of the same age. Male users aged between 35 to 44 years old accounted for 9.8% of the total YouTube audience, while the female audience of the same age group accounted for 8.1%.
Below we have mentioned a table showcasing YouTube users worldwide by Age group and Gender as of April 2024:
According to reports, the percentage of male audience accessing YouTube was 54.3% and the percentage of female audience accessing the platform was 45.7% which is slightly lower compared to males in April 2024.
As of April 2024, India was the leading country with the largest number of YouTube audiences worldwide with around 476 million users engaging with YouTube. The United States is ranked second in this list with 238 YouTube viewers, while Brazil takes third place with 147 million YouTube viewers.
Below we have mentioned a table showcasing YouTube audience size by Country:
In 2023, users globally spent about 28 hours accessing YouTube mobile apps per month. Users in South Korea became the leading country with the highest monthly time spent engaging with video content on YouTube apps by 45 hours and 49 minutes. Followed by Thailand in second position with 42 hours and 29 minutes. Russia is ranked in third position with 37 hours and 25 minutes.
Below we have mentioned a table showcasing the monthly time spent on the YouTube mobile app per user as of December 2023:
Average Time Spent on YouTube in the US (2019 to 2024)
According to the latest reports, the average time spent by users on YouTube in the US has witnessed continuous growth over the years. The United States is the second country falling behind India with the highest number of YouTube users worldwide. Currently, the US has 238 million users on the platform.
In 2019 the average time spent by a user on YouTube was recorded at 39.7 minutes and ever since the platform has seen nothing but constant growth in the number of average time spent by a user. In 2020, the average time reached 43.7 minutes, while in 2021 it reached 45 minutes.
A total of 10.1% of the rise in the average time was seen between 2019 to 2024. Some experts suggest that the global pandemic has played a significant role in the rise of social media usage among users, which has impacted the growth of YouTube’s average time spent.
In 2022, it also saw a growth of 2.7% and the numbers reached 46.2 minutes, while a rise of 2.8% was again seen in 2023 which increased the average time to 47.5 minutes.
In 2024, United States users are projected to see a rise in YouTube’s watch time by 2.5% resulting in the average time spent reaching 48.7 minutes.
India uses YouTube the most and has 476 million users, followed by the United States in the second position with 238 million users accessing the platform.
How many users are currently on YouTube?
Currently, YouTube has 2.9 billion monthly active users accessing the platform. Along with this, various reports also suggest the number of monthly active users on YouTube will reach 3.1 billion in 2025.
Is YouTube still growing?
Yes, YouTube is still growing and witnessing excellent growth every day. In fact, the global number of YouTube users is expected to see a rise in its users by 232.5 million between 2024 to 2029.
How many users use YouTube in India?
Around 476 million users use YouTube in India, and India is the leading country with the highest number of YouTube users than any other country.
Do YouTubers have a future?
Yes, looking at the current statistics YouTube seems to have a good future with an excellent rise in its user base from all across the world.
Wrapping Up
Overall, YouTube has an excellent user base in 2024 with 2.9 billion monthly active users. YouTube has placed its dominance in the online video-sharing market for people from different age groups and geographical locations offering a wide range of video content for a variety of purposes such as news, entertainment, learning & education, etc.
Looking at the current statistics of YouTube, the platform is expected to keep evolving with continuous growth in the future.
OpenAI released its latest multimodal large language model “GPT-4” last year on 14th March 2023. The advanced LLM GPT-4 has been making waves globally ever since its release. This powerful language model offers excellent capabilities such as text generation, image recognition, translation, code writing, and more.
In this article, we are going to dive deep into GPT-4 statistics and understand its processes and trends. We will also include the number of parameters GPT-4 is trained on, the performance of GPT-4 on standardized tests, benchmark comparisons, and more.
Top GPT-4 Statistics
GPT-4 was released by OpenAI on 14 March 2023.
GPT-4 is capable of processing and generating inputs and outputs in various formats such as text, image, audio, and video.
The response time to audio input in GPT-4 can go as little as 232 milliseconds, with an average of 320 milliseconds.
GPT-4 is capable of Achieving 88.7% on 0-shot COT MMLU for general knowledge questions.
The context length of GPT-4 is 8,192 tokens.
GPT-4 Turbo is capable of considering about 96,000 words within one go.
OpenAI’s GPT-4 can outperform Whisper-v3 in speech recognition and translation.
GPT-4 is the latest large language model developed by OpenAI, the organization that created the popular AI chatbot ChatGPT. This LLM is known for its impressive capabilities of text creation, code generation, and translation.
GPT-4 is a Multimodal LMM
The stand-out quality of GPT-4 that sets the large language model apart from its predecessors GPT-3 is its multimodality. This means that GPT-4 is capable of processing and responding to data and information in numerous formats such as:
Text (written language)
Images (visual information)
Code (programming language)
Audio
Was GPT-4 Really Trained on 100 Trillion Parameters?
No, GPT-4 wasn’t trained on 100 trillion parameters. OpenAI CEO Sam Altman claimed the rumor is untrue.
While other sources claim GPT -4 has close to 1.8 trillion parameters across 120 layers, this makes it almost ten times larger than its previous LLM, GPT-3, which had around 175 billion parameters.
ChatGPT Plus Subscribers (GPT-4 User Numbers)
There are around 180.5 million active users on ChatGPT. An estimated 1% of subscribers have subscribed to ChatGPT Plus, which gives them access to GPT-4 for a monthly price of $20. This makes the number of users around 180,000 to 200,000.
With a context length of around 8,192 tokens. OpenAI has started to provide limited access to 32,768 content (which is the text of about 50 pages) version GPT-4-32k. Previously, GPT-4 had featured a content window of 8,000 tokens with a 32K model available for some developers via API. However, this context window has now been surpassed for turbo models. Therefore, this indicates that GPT-4 Turbo is capable of considering about 96,000 words in a single go which is higher than novels.
In this section, we are going to compare GPT-4 with GPT-3 to understand its true capability and how it performs in terms of generating relevant and accurate answers:
Accuracy and Relevance: In terms of providing more relevant and precise answers to users’ questions. Studies have found GPT-4 to be a more capable large language model that can generate answers with an accuracy of 85.5% in a 3-shot massive multitask language understating. GPT-3.5 had an accuracy rate of 70.1% which is lower than GPT-4.
Image Recognition: OpenAI’s latest LLM GPT-4 can see, analyze, and describe images with complete detail. For example, if you show an image of a particular food item, GPT-4 can instantly suggest recipes for you. GPT-3 on the other hand, doesn’t have an image recognition feature available.
Audio and Vision: ChatGPT’s premium plan known as “ChatGPT Plus” allows users to engage in conversation with GPT-4 through its voice control feature which is available on the mobile app. It also allows users to effortlessly explain and describe things to the GPT via drawing.
Mathematical Skills: Studies suggest that GPT-4 is much more capable of solving mathematical problems and generating accurate and relevant answers to users’ questions compared to GPT-3.
Language: The language test revealed that GPT-4 has outperformed GPT-3.5 in 24 out of 26 tests, including low-resource languages such as Welsh, Latvian, and Swahili. In addition, GPT-4 also outperformed GPT-3.5 in paraphrasing, text edition, and solving grammatical errors.
Safety Measures: When it comes to privacy and security measures, GPT-4 is better than GPT-3.5 in avoiding sensitive, disallowed, and inappropriate requests. Apparently, OpenAI has decreased its model’s tendency to respond to disallowed content by 82% and improved its response to sensitive requests by 29%.
Benchmark comparison between Google’s Gemini Ultra and OpenAI’s GPT-4V in 2024
A benchmark comparison was made between OpenAI’s GPT-4 and Google’s Gemini Ultra to understand its true capability and performance. Gemini Ultra has beaten OpenAI’s GPT-4 in the MMMU benchmark. The only two categories where GPT-4 performed better than Gemini Ultra were Business and Science, while the overall quality of the models was pretty similar.
Below we have mentioned a table showcasing the benchmark comparison of Google’s Gemini Ultra and OpenAI’s GPT-4:
Code Benchmark comparison between Gemini Ultra and GPT-4 in 2024
OpenAI’s GPT-4 and Google’s Gemini Ultra are currently some of the most popular generative AI platforms globally, utilized by users worldwide. Surprisingly, both GPT-4 and Gemini Ultra have quite similarities in the code creation benchmark. Gemini Ultra is a bit ahead of GPT-4, beating it in both the coding benchmark and MMLU benchmarking.
Below we have mentioned a table showcasing a code benchmark comparison between Gemini Ultra and GPT-4 in 2024:
In this section we are going to take a look at some of the top commercial uses of GPT-4:
1. Duolingo: A Language Learning
Popular language learning app Duolingo turned towards OpenAI’s newest large language model GPT-4 to enhance its capabilities by adding two new features: A “Roleplay” feature which acts as an AI conversation partner and A “Explain my Answer” feature that breaks down the rules and alerts the user when they make a mistake along with the reasoning. Users can access both of these features through Duolingo’s new subscription tier known as Duolingo Max. By adding the latest features, users will be able to learn from their mistakes as GPT-4 generates initial responses by simply clicking on “Explain my answer.” From there, users can directly return to the lesson or gain further in-depth explanations. (Source)
2. Be My Eyes, Visual Impairment Assistant
The Danish mobile app Be My Eyes is a visual impairment assistant specially built to help blind or visually impaired users recognize elements and objects. Be My Eyes collaborated with OpenAI in March 2023 to create a new tool called “Be My AI” that can describe the visual world to blind or visually challenged people. This new product was incorporated with OpenAI’s latest LLM GPT-4V which assists in providing a description of the images taken on the smartphone by the blind user. “Be My AI” was piloted from March 2023 to August 2023 with approximately 200 blind or low-vision beta testers to assess the safety and user experience of the tool and point out if the platform required any room for improvement or changes. By September 2023, the group of beta testers had reached 16,000 blind or visually impaired users who were requesting almost 25,000 descriptions a day. (Source)
Be My Eyes also followed it by the development of a virtual volunteer which is also powered by OpenAI’s GPT-4 available within the app of Be My Eyes that can create the same level of understanding and context as a human volunteer. ?(Source)
3. Stripe, Fraud Detection
The American multinational financial services and software, Stripe has leveraged OpenAI’s GPT-4 to fight fraud and enhance its user experience. Previously, Stripe has utilized GPT-3 to ensure a good support team and help its users through various tasks such as summarization of a user question and routing issue tickets. But with GPT-4 Stripe is able to perform various additional tasks such as scanning sites and providing a short summary of the available data, which apparently outperformed human-written summaries.
It was revealed that when performing a hand-checking on the results it was found that humans were wrong and in fact summaries generated by the model were correct. The Product Lead from Stripe stated, “GPT-4 was better compared to human reviewers.”
Another excellent capability of Stripe includes its extensive technical documentation and robust developer support team that can effortlessly answer all the troubleshooting issues or technical questions ?instantly. GPT-4 can even scan the inbound communications and recognize any inappropriate or coordinated activity.
Iceland has collaborated with OpenAI for the utilization of its advanced LLM GPT-4 along with the help of private industry to preserve the Icelandic language and to transform the defensive position into finding new ways to innovate. The core reason behind this collaboration was not only to boost the capability of GPT-4 to offer service to a new corner of the world but also to generate resources that truly promote the preservation of various low-resource languages.
In the Reinforcement learning from Human Feedback, human testers provided a prompt to GPT-4 and it reverted back with four possible answers to the tester’s input. Once the answers are generated the testers select the best possible answer from the available options and further edit it to generate the best output. The data from this process is then utilized to train GPT-4 to generate improved responses in the future. (Source)
5. Khan Academy, Education
Khan Academy made an announcement in March 2023, that it will utilize GPT-4 to power Khanmigo which is an AI-powered assistant that operates both a virtual tutor for students along a classroom assistant for teachers. (Source)
The non-profit organization began its testing process with OpenAI’s latest language model back in 2022 and initially started by making the Khanmigo pilot program specially available for a limited number of users, while others can join the waitlist. Apart from this, Khan Academy has also started the adoption of GPT-4 for teachers. It has begun the testing process through which teachers can utilize GPT-4 for different purposes such as writing classroom prompts or generating instructional materials.
6. Morgan Stanley, Wealth Management
Morgan Stanley has leveraged GPT-4 to arrange and sort out its extensive knowledge base.
At the beginning of last year, the company started to explore ways to harness the intellectual capital with GPT’s retrieval and embedding capabilities first with the use of GPT-3 and then GPT-4. ?
Morgan Stanley has utilized GPT-4 to make its internal chatbot more helpful for the company’s needs. Today, over 200 employees are querying the system every day and providing their feedback effortlessly.
FAQs
Can I use GPT-4 for free?
Yes, users can use GPT-4 Turbo for free in Microsoft Pilot and receive accurate and better results by accessing the latest AI chatbot capabilities.
What app uses GPT-4?
Various apps utilize GPT-4, such as the language learning app Duolingo, the Visual Impairment Assistant Be My Eyes, and more.
Can GPT-4 do coding?
Yes, GPT-4 can perform coding. GPT-4 can even assist developers by generating various useful suggestions, creating code snippets, and offering various insights on coding. It can be utilized as a complement to human coding skills instead of acting as a replacement.
Can GPT-4 generate images?
Yes, GPT-4 is capable of generating images. It’s a multimodal large language model, and therefore, it has capabilities including text, image, audio, and video outputs.
Wrapping Up
GPT-4 is a powerful and versatile language model and its immense scale, safety features, and capability to handle different forms of data have increased its demand among users across the world.
But GPT-4 is still relatively new and with higher accessibility and functionality the influence of the LLM is expected to grow. From content generation to coding to summarization to image recognition GPT-4 has the potential to transform the way we interact with technology and data.
The H-1B visa program plays an important role in the U.S. workforce by allowing employers to hire highly skilled foreign professionals, particularly in specialized fields such as technology, healthcare, engineering, and finance.
Each year, thousands of workers receive new visas or renew existing ones, contributing significantly to innovation, research, and economic growth. In 2025, approximately 479,953 registrations were submitted.
This article provides a detailed overview of H-1B visa statistics, including total registrations, lottery selections, income levels across professions, and more.
What is an H1B Visa?
The H1B visa is a U.S. non-immigrant visa that allows U.S. employers to hire foreign workers in “speciality occupations temporarily.” These are jobs that require a bachelor’s degree or its equivalent in a specific field. It’s a key part of the U.S. immigration system, particularly for the technology and IT sectors.
H1B visas are designed to help U.S. companies fill workforce gaps with highly skilled foreign talent.
Initially granted for up to three years and extendable to a maximum of six years.
Further extensions may be possible if the worker is in the process of applying for a green card.
H1B Visa Registrations (2021-2025)
H1B visa registrations have shown notable fluctuations in recent years. Total registrations increased steadily from 274,235 in FY 2021 to a peak of 780,342 in FY 2024, reflecting growing demand for skilled foreign workers.
In FY 2025, registrations declined to 479,953, indicating a moderation in applicant numbers following the previous year’s surge. This trend highlights both the high demand for H1B visas and the year-to-year variability in the application process.
Cap Fiscal Year
Total Registrations
2021
274,235
2022
308,613
2023
483,927
2024
780,342
2025
479,953
H1B Visa Selection (2020-2025)
The H-1B visa has become more competitive in recent years due to growing demand and limited availability. The number of visas selected each year has stayed relatively steady, with 124,415 in 2021, 131,924 in 2022, 127,600 in 2023, a peak of 188,400 in 2024, and then 135,137 in 2025.
This shows that while interest in the H-1B program continues to rise, the chances of being selected remain low, making the process highly competitive for applicants across degree types.
Cap Fiscal Year
Total H1B Selections in Lottery Process
2021
131,924
2022
127,600
2023
188,400
2024
135,137
2025
120,141
H1B Visas Approved by Country
In 2024, the majority of H1B visa recipients came from India, which accounted for 71% (283,397) of the total, reflecting the country’s strong presence in the U.S. technology and IT workforce. China followed with 11.7% (46,680), while other countries such as the Philippines (1.3%), Canada (1.1%), South Korea (1.0%),Mexico (0.8%), Taiwan (0.8%), and Pakistan (0.8%) made up smaller but notable shares.
This highlights how the H1B program continues to serve as the primary pathway for highly skilled professionals, particularly Indian tech talent, to work in the United States.
Countries
H1B Visa Recipients
India
283,397
China
46,680
Philippines
5,248
Canada
4,222
South Korea
3,983
Mexico
3,333
Taiwan
3,099
Pakistan
3,052
Brazil
2,641
Nigeria
2,275
Nepal
2,157
United Kingdom
1,727
Colombia
1,646
Turkey
1,432
France
1,281
Iran
1,237
Vietnam
1,213
Bangladesh
1,173
Spain
1,043
Germany
1,030
H1B Recipients by Occupation
In FY 2024, more than six in ten H1B visa recipients were employed in computer-related occupations, making up 63.9% of all approved beneficiaries. Other major fields included architecture, engineering, and surveying(10.2%), education (6.0%), and administrative specialization (5.4%).
Smaller shares went to medicine and health (4.2%), mathematics and physical science (2.8%), life sciences (1.9%), and managerial roles (1.7%). This breakdown highlights the technology sector’s dominance in driving H1B demand, while also demonstrating its significance across various professional fields in the U.S. workforce.
Top 10 employers of H1B visa holders as of June 2025
As of June 2025, the top employers of H1B visa holders include a mix of technology giants, consulting firms, financial institutions, and retailers. Amazon.com Services leads the list with over 10,000 H1B workers, followed by Tata Consultancy Services, Microsoft, and Meta, each employing thousands of skilled professionals.
Other major employers include Apple, Google, Cognizant, JPMorgan Chase, Walmart, and Deloitte, highlighting how H1B talent is spread across sectors ranging from IT and consulting to finance and retail. This demonstrates the continued reliance of U.S. companies on global talent to fill critical skill gaps.
Companies
H1B Visa Holders
Amazon.com Services LLC
10,044
Tata Consultancy Services Limited
5,505
Microsoft Corporation
5,189
Meta Platforms INC
5,123
Apple INC
4,202
Google LLC
4,181
Cognizant Tech Solutions US Corp
2,493
JPMorgan Chase and Co
2,440
Wal-Mart Associates INC
2,390
Deloitte Consulting LLP
2,353
Median Income for New H1B Visa Workers
New H1B visa approvals in 2025 show a wide range of median incomes across occupations, reflecting both demand and specialisation. Computer-related roles make up the largest share, with over 75,000 new visas and a median income of $101K, while professional, technical, and managerial workers earn among the highest at $125K. Law and jurisprudence positions top the chart with a striking $207K median income.
Other high-paying areas include managers and officials ($110K) and math and physical sciences ($104K). In contrast, fields like education ($68K), entertainment and recreation ($64K), and religion and theology ($49K) report lower median earnings, highlighting significant variation by profession.
Occupation
New Visa’s
Median Income (New Visa)
Computer-related
75,428
$101K
Architecture, Engineering and Surveying
15,123
$97K
Education
13,725
$68K
Administrative specializations
9,270
$94K
Medicine and Health
8,492
$80K
Math and Physical Sciences
4,781
$104K
Life Sciences
3,927
$77K
Managers and officials
2,589
$110K
Professional, technical and managerial
2,345
$125K
Social Sciences
1,806
$97K
Unknown
1,202
$82K
Law and Jurisprudence
946
$207K
Art
813
$80K
Museum, Library and Archival Sciences
197
$93K
Miscellaneous
224
$53K
Writing
145
$71K
Entertainment and Recreation
94
$64K
Religion and Theology
93
$49K
Sales and Promotions
5
$78K
Median Income for Renewed H1B Visa Workers
Renewed H1B visa workers in 2025 earn notably higher median incomes compared to new approvals, reflecting career progression and experience. Computer-related occupations, with nearly 180,000 renewed visas, report a median salary of $135K, while high-demand fields like medicine and health ($141K), math and physical sciences ($140K), and education ($140K) also see strong earnings.
Professional, technical, and managerial roles reach one of the highest medians at $162K, with law and jurisprudence leading all fields at $230K. By contrast, administrative specializations ($76K), entertainment ($74K), and miscellaneous roles ($72K) remain on the lower end, underscoring disparities across professions.
Occupation
Renewed Visas
Median Income (Renewed Visas)
Computer-related
179,822
$135K
Architecture, Engineering and Surveying
25,546
$126K
Education
10,053
$140K
Administrative specializations
12,399
$76K
Medicine and Health
8,445
$141K
Math and Physical Sciences
6,465
$140K
Life Sciences
3,466
$100K
Managers and officials
4,033
$155K
Professional, technical and managerial
2,554
$162K
Social Sciences
1,579
$130K
Unknown
1,497
$124K
Law and Jurisprudence
873
$230K
Art
930
$112K
Museum, Library and Archival Sciences
200
$145K
Miscellaneous
113
$72K
Writing
125
$92K
Entertainment and Recreation
53
$74K
Religion and Theology
30
$81K
Sales and Promotions
7
$84K
H1B Visa Approval Rate by Gender
In FY2022, H-1B visa approvals showed a clear gender gap, with male applicants accounting for 70.8% of approvals compared to just 29% for female applicants. This highlights the continued male dominance in H-1B visa grants, reflecting broader trends in the technology and STEM fields where most H-1B workers are employed.
Gender
Approval Rate
Male
70.8%
Female
29%
Largest number of H1B Recipients by City
The largest number of H1B visa recipients in the United States is concentrated in a few key metropolitan areas. New York City leads with 372,100 approvals, followed by San Jose with 215,700, San Francisco with 165,000, and Dallas with 150,200.
These cities collectively represent major hubs for technology, business, and innovation, reflecting the significant role of H1B workers in driving economic and technological growth in these regions.
City
Number of H1B Recipients
New York City
372,100
San Jose
215,700
San Francisco
165,000
Dallas
150,200
Type of Approved H1B Petitions
In FY2024, approved H1B petitions were distributed across both new and continuing employment categories. Petitions for new employment accounted for 35%, while new concurrent employment made up less than 1%.
Among continuing employment, extensions of stay represented 25%, amendments 24%, and change of employer (sequential employment) 16%. This breakdown highlights that while a significant portion of approvals support new hires, a substantial share also facilitates ongoing employment adjustments for existing H1B workers.
Type of H1B Petitions
Approved Petitions
New Employment
35%
Extension of Stay
25%
Amendment
24%
Change of Employer
16%
H1B Visa Approval Rate
The H-1B visa approval rate has remained consistently high over the past five years, averaging around 97–98%. In 2025, the approval rate stood at 97.8%, with 9% of cases completed through a Request for Evidence (RFE) and 85.4% approved after an RFE.
Similar trends were seen in previous years, with 98% approvals in both 2022 and 2024, and slightly lower at 97.2% in 2023 and 97.3% in 2021. While RFEs remain a part of the process, the data show that the majority of applicants who receive them are still able to secure approval.
Fiscal Year
H1B Approval
Completion with RFE %
Approved with RFE %
2025
97.8%
9.0%
85.4%
2024
98%
8.2%
84.9%
2023
97.2%
9.9%
81.0%
2022
98%
9.6%
85.5%
2021
97.3%
16.2%
86.5%
H1B Visa Denial Rate
The denial rate for initial H1B petitions has experienced significant fluctuations over the past decade. It climbed from 6% in FY 2015 to a peak of 24% in FY 2018 and FY 2019, reflecting stricter adjudication during that period.
However, the trend reversed in subsequent years, with steady improvements leading to a record low denial rate of just 2.15% in FY 2025. This decline highlights a more favourable environment for H1B applicants in recent years.
Fiscal Year
Denial Rate
2025
2.15%
2024
2.5%
2023
3.5%
2022
2.2%
2021
4%
2020
13%
2019
21%
2018
24%
2017
13%
2016
10%
2015
6%
Top Reasons for H1B Visa Denial
A successful H1B application requires both the employer and the employee to meet all visa requirements. Any errors or ineligibility at any of the three stages can result in denial. Some of the most common reasons include:
Not meeting the speciality occupation requirements.
Lack of appropriate academic qualifications.
Failure to provide proof for third-party work-site arrangements.
Improper maintenance of current residency status.
Inability to establish a valid employer-employee relationship.
Employee being deemed inadmissible to the U.S.
Inability to demonstrate the ability to pay the prevailing wage.
Not responding to a Request for Evidence (RFE).
Failure to pay the required filing fees.
H-1B Visa Chances for Bachelor’s Degree Graduates
For applicants with a bachelor’s degree, the chances of securing an H-1B visa have varied significantly over the years. In 2020, the acceptance rate was 21.6%, but it dropped steadily to just 7.67% in 2023, the lowest point in recent years.
The rate improved slightly in 2024 to 12.37% and rose further to 16.92% in 2025. Overall, the data show that while opportunities for bachelor’s degree graduates remain limited, the chances have seen a modest recovery after a sharp decline.
Year
H1B Visa Acceptance
2020
21.6%
2021
19.31%
2022
12.27%
2023
7.67%
2024
12.37%
2025
16.92%
Recent H1B Visa Policy Changes
A recent proclamation by the Trump administration has introduced a substantial new fee, which has the potential to reshape the H1B landscape significantly.
As of September 21, 2025, a new $100,000 fee is required for certain new H1B petitions. This fee is meant to be paid by the sponsoring company, not the applicant. The purpose of this fee is to deter companies from over-relying on foreign workers and to generate revenue for domestic training programs.
Impact of H1B Visa Fee Increase
The dramatic fee hike is expected to lead to a significant statistical decline in new H1B visa applications.
Reduced Application Volume: Analysts at JPMorgan Chase estimate a potential reduction of up to 5,500 new work authorizations per month. This would drastically lower the number of H1B visas issued, possibly by over 60,000 per year, assuming the demand for the visas isn’t completely eliminated.
Shift to Alternative Visas: Companies and highly skilled professionals are expected to pivot to alternative visa categories with more favorable cost structures and no annual cap.
L-1 Visa: This visa for intracompany transfers is now a much more attractive option, especially for multinational corporations that can transfer employees from their overseas offices. The average cost for an L-1 visa is around $7,000, making it a dramatically cheaper alternative.
O-1 Visa: The O-1 visa for individuals of “extraordinary ability” is also gaining traction. With an average cost of around $12,000 and no annual cap, it offers a pathway for top-tier talent. The statistical approval rate for O-1 visas has been higher than H1B visas, at around 93% vs. the H1B’s pre-2025 rejection rate of up to 24%.
Future Outlook for H1B Visa
The future of H-1B visas is expected to be more competitive and selective. A decrease in the total number of applications is expected due to the new, significant application fees that have been implemented.
This could also lead to a shift in the type of applicants, with a greater focus on highly skilled, higher-paid workers as the U.S. government moves towards a more merit-based system. Consequently, smaller companies may find it more difficult to hire foreign talent, potentially reducing their reliance on the H-1B program.
Additionally, the lingering green card backlogs and stricter visa scrutiny could encourage skilled professionals to seek opportunities in other countries with more straightforward immigration pathways.
Cord cutting has become one of the biggest changes in how people watch TV over the past ten decade, as millions of households move away from traditional cable and satellite TV in favor of streaming services, over-the-air (OTA) channels, and other digital alternatives.
Driven by factors such as high subscription costs, the convenience of on-demand content, and the rise of popular streaming platforms like Netflix, Hulu, and Disney+, cord cutting is reshaping how Americans and viewers worldwide consume entertainment.
In this article, we are going to take a look at 30+ cord cutting statistics showcasing how viewers are moving away from traditional TV, why they are making the switch, and more.
Top Cord Cutting Statistics
U.S. Cord-Cutting Households Projected to Reach 60.3 Million by 2027
The number of cord-cutter households in the United States is projected to grow steadily through the middle of the decade, reflecting the continued shift away from traditional pay TV services. In 2020, approximately 30.8 million U.S. households had cut the cord, a figure that increased to 36.5 million in 2021 and reached 47.6 million by 2023.
This upward trend is expected to continue, with cord-cutter households forecast to rise to 51.5 million in 2024, 54.7 million in 2025, and 57.6 million in 2026. By 2027, the total number of U.S. cord-cutting households is projected to reach 60.3 million, underscoring the accelerating decline of traditional pay TV adoption.
Year
U.S. Cord-Cutter Households
2027
60.3 million
2026
57.6 million
2025
54.7 million
2024
51.5 million
2023
47.6 million
2022
43 million
2021
36.5 million
2020
30.8 million
Cord-Cutting Drives Non-Pay TV Households to 80.7 Million by 2026
U.S. cord-cutting is projected to accelerate sharply over the coming years, signaling a major transformation in household television consumption. By 2026, an estimated 80.7 million U.S. households are expected to be non-pay TV households, compared with just 54.3 million households still subscribed to traditional pay TV services.
This represents a dramatic reversal from 2018, when pay TV dominated with 90.3 million subscribing households, while only 37.3 million households relied on non-pay TV options. The gap has steadily narrowed year over year, with non-pay TV households projected to overtake pay TV households by 2026.
Year
Pay TV Households
Non-Pay TV Households
2026
54.3 million
80.7 million
2025
56.8 million
77.2 million
2024
59.6 million
73.2 million
2023
62.8 million
68.9 million
2022
66.4 million
64.3 million
2021
71.6 million
58.3 million
2020
77.5 million
50.9 million
2019
84.4 million
44.6 million
2018
90.3 million
37.3 million
Cord-Cutting Pushes Non-Traditional TV Households to 75% by 2026
The share of U.S. households with a television but without a traditional TV subscription has risen steadily over the past decade and is expected to reach a critical threshold by 2026. In 2015, only 20% of TV households lacked a traditional cable or satellite connection.
By 2023, that share had climbed to over 60%, representing a 7 percentage point increase from the previous year alone. This upward trend is projected to continue, with an estimated 75% of U.S. TV households expected to be without a traditional TV subscription by the end of 2026.
Cord-Cutting Trends and U.S. TV Subscription Statistics
35 Million Americans Have Never Subscribed to Cable Television
Around 35 million Americans have never had a cable TV subscription, showing a clear shift in viewing habits among younger consumers. Many of these individuals are younger adults, including college students and recent graduates, who grew up using streaming services instead of traditional cable. As they start their own households, they are choosing streaming platforms over cable, contributing to the long-term decline of traditional pay TV adoption.
U.S. Pay TV Has Lost Over 20 Million Subscribers Since 2014
U.S. cable and satellite television providers have experienced a significant and sustained decline in subscribers over the past decade. Since 2014, pay TV services have lost more than 20 millionsubscribers, highlighting a long-term shift away from traditional television platforms.
The pace of subscriber losses has accelerated in recent years, with the COVID-19 pandemic acting as a major turning point. In 2020 alone, more than 5 million U.S. households canceled their pay TV subscriptions, marking one of the steepest annual declines on record.
Pay TV Adoption Among U.S. Adults Falls From 76% to 56% (2015 to 2021)
The share of American adults subscribing to cable or satellite television has declined sharply over the past decade. In 2015, approximately 76% of U.S. adults paid for a traditional pay TV package. By 2021, that figure had fallen to just 56%, representing a 26% drop in adoption over six years.
Cable TV Prices Reach $217 per Month, Surpassing Household Utilities
The rising cost of traditional cable television has become a major factor driving cord-cutting in the U.S. The average household cable package now costs approximately $217.42 per month, making it more expensive than the combined average monthly cost of all other household utilities, which totals about $205.50.
Comcast Video Subscribers Fall From 22.6M to 14.5M Since 2014
Cord cutting has had a substantial impact on major U.S. pay TV providers, including industry leaders. Comcast, the largest cable TV operator in the United States, has lost more than 8 million video subscribers since 2014, with its subscriber base declining from 22.6 million in 2014 to approximately 14.5 million today. A similar trend is evident in the satellite TV segment, as DirecTV has shed around 8.42 million subscribers over the same period.
Over One-Third of Americans Maintain Both Cable and Streaming Subscriptions
Despite the rapid growth of streaming, a significant share of U.S. consumers continues to maintain a mix of traditional and digital viewing options. Survey data shows that over one-third of Americans have both cable and streaming subscriptions, with 32% reporting that they currently pay for cable in addition to streaming services such as Netflix and Hulu.
A further 3% indicate they previously relied only on streaming but have since added a cable subscription. At the same time, 11% of Americans say they have exclusively used streaming services, highlighting the coexistence of hybrid viewing habits alongside fully digital consumption.
Live Sports Drives 18% of Former Cord-Cutters Back to Cable
Recent research suggests that cord-cutting is not always permanent for a large share of consumers. A survey of 500 cable TV subscribers found that nearly half are former cord-cutters who returned to traditional cable services after previously canceling them.
Among those who resubscribed, access to live sports was the leading reason, cited by 18.4% of respondents, followed by live entertainment events at 10.0%. Additionally, 9.7% reported that bundled offerings, including cable packaged with internet or other services, were a key factor in their decision to return, highlighting the continued appeal of live content and bundled value propositions.
23% of Cord-Cutters Turn to Free Over-the-Air TV
About 23% of people cut the cord to watch free TV using an antenna. Antennas usually cost between $50 and $200 as a one-time purchase, with no monthly fees. This setup allows viewers to access local channels and some live sports for free, making over-the-air TV a simple and low-cost option for casual viewers.
High Costs Drive Cord-Cutting for 86.7% of Consumers
Cost remains the dominant factor driving cord cutting among U.S. consumers. Survey data shows that 86.7% of respondents cite high prices as the primary reason for canceling traditional pay TV services. Changing viewing preferences also play a significant role, with 39.7% indicating a preference for streaming platforms over cable or satellite TV.
Additionally, 23% of respondents report switching to antenna-based TV as an alternative, while 15.9% favor binge-watching formats that are more common on streaming services. Lifestyle changes also contribute, as 13% cite moving and choosing not to renew services, and 7.7% point to a preference for original content available on streaming platforms.
Top Reasons
Share of respondents
Price
86.7%
Preference for streaming
39.7%
Switching to antennae TV
23%
Prefer binge watching
15.9%
Moving and don’t want to renew services
13%
Preference for original content on streaming services
7.7%
Young Adults Lead U.S. Cord-Cutting, Cable Subscriptions Fall to 34%
Young adults are leading the cord-cutting trend in the U.S. Among 18 to 29 year olds, the share with a cable TV subscription dropped from 65% in 2015 to 34% in 2021, the largest decline across all age groups.
Adults aged 30 to 49 also reduced their cable usage, falling from 73% to 46% over the same period. In contrast, Baby Boomers remain the most loyal subscribers, with 86% of those 65 and older having cable in 2015 and 81% still subscribed today.
52% of Former Cable Subscribers Report No Regret After Cord-Cutting
Most U.S. households are satisfied after cutting the cord. Survey results show that 52% of former cable subscribers say they do not miss anything after canceling their service. Among those who do miss certain content, 23% say they miss live sports and 22% miss local or national news. However, many of these viewers are able to access sports and news through live TV streaming services or over-the-air (OTA) options, reducing the need to return to traditional cable TV.
30% of U.S. Cable Subscribers Consider Canceling Their Service
About 30% of Americans with cable TV say they are likely to cancel their subscriptions. Among the 45% of adults who had cable in February 2022, 21% said they were likely to cancel, and 9% said they were very likely to do so. Meanwhile, 25% were neutral, 11% were unlikely, and 34% said they were very unlikely to cancel, showing that a significant portion of cable subscribers are considering leaving.
Only 5% of Cord-Cutters Regret Canceling Cable or Satellite TV
Cord-cutting satisfaction remains very high among U.S. consumers, with only about 5% of former cable or satellite TV subscribers expressing regret over canceling their traditional TV services. This low level of dissatisfaction highlights that the vast majority of cord-cutters are content with their decision and find alternative viewing options, such as streaming or over-the-air TV, sufficient for their entertainment needs.
Live Events Drive 29% of Former Cord-Cutters Back to Pay TV
Over one-third of U.S. and Canadian cord-cutters resubscribed to pay TV services because they were unable to access all the entertainment they wanted. Cost was less of a concern for many, as over 2 in 10 respondents said the services were not very expensive. Additionally, 29.1% of those who returned to pay TV indicated that it was the best way to watch major live events.
Reasons
Percentage of Subscribers
Couldn’t get all the entertainment they wanted
33.8%
Consider it the best way to watch major live events
29.1%
Consider it the best way to watch local programming
27.5%
Consider it the best way to watch sports
27.1%
like knowing they always have many channels available
25.8%
have entertainment components that only work with pay TV
23.2%
Think that service is not very expensive
19.7%
Streaming video services are not as reliable
18.8%
got TV service bundled with my broadband Internet service
17%
They can easily skip the ads when recording programs from pay TV
17%
89% of Cord-Cutters Would Have Cancelled Cable Regardless of COVID-19
The COVID-19 pandemic had only a limited effect on cord-cutting decisions. According to Media Play News, 89% of recent cord-cutters said they would have canceled their cable subscriptions even without the pandemic.
Early in the pandemic, in July 2020, 31% of cord-cutters said they wouldn’t have canceled if not for COVID-19, but by February 2021, that number dropped to just 11%, showing that most cancellations were driven by long-term trends rather than the pandemic.
Streaming Adoption Trends
U.S. Streaming Subscriptions Reach 354 Million Nationwide
The U.S. streaming market has reached a substantial scale, with an estimated 354 million active streaming subscriptions nationwide. Streaming adoption continues to expand rapidly, reflecting strong consumer demand for on-demand and flexible viewing options. By 2027, streaming penetration is projected to reach 86% of U.S. households with a television, totaling approximately 458 million subscriptions.
Global SVOD Subscriptions to Reach 1.68 Billion by 2027
Global streaming video-on-demand (SVOD) subscriptions are projected to reach 1.68 billion by 2027, representing an increase of 475 million from 2021, according to Digital TV Research.
The leading platforms in terms of subscriptions are expected to be Netflix with 251 million, Amazon Prime Video with 250 million, Disney+ with 207 million, Paramount+ with 82 million, and HBO Max with 77 million. These projections highlight the continued rapid growth of the global streaming market and the dominance of major players in the SVOD space.
Hulu, Netflix, and Amazon Prime Provide Vast Entertainment Library Under $40/Month
With just three streaming subscriptions Hulu, Netflix, and Amazon Prime Video cord-cutters can access over 11,000 seasons of TV shows and more than 13,000 movies, according to Ampere Analysis. All three services together cost less than $40 per month, providing a vast library of commercial-free entertainment at a fraction of the cost of traditional cable.
Netflix Revenue Rises From $3B to $29.7B in a Decade
Netflix has experienced remarkable growth over the past decade, with annual revenue increasing from just over $3 billion ten years ago to $29.7 billion today. The company reported a net income of $5.16 billion and employs more than 11,000 people across 60 countries. This growth reflects Netflix’s global expansion and its position as a leading streaming service worldwide.
Netflix Leads U.S. Streaming with 60% Weekly Viewership
Netflix continues to be the most popular streaming platform in the U.S., with about 3 in 5 Americans (60%) watching it at least once per week. Other widely watched platforms include YouTube at 44%, Hulu at 42%, Amazon Prime Video at 39%, and Disney+ at 27%, showing strong competition but clear dominance by Netflix in weekly viewership.
89% of Marketers Say Streaming Ads Are as Effective as Traditional TV
Marketers are increasingly prioritizing streaming platforms as a core advertising channel, reflecting changing audience behavior and media consumption patterns. According to industry data, 89% of marketers believe that advertising on streaming platforms is equally or more effective than traditional television advertising.
This confidence is translating into budget shifts, with 43% of marketers now ranking streaming services as the top channel for brand storytelling. By comparison, only 29% place social media first, while 26% still consider traditional TV their primary storytelling medium, underscoring the growing dominance of over-the-top (OTT) platforms in modern advertising strategies.
Apple TV+ Reaches 25 Million Paid Subscribers
Apple TV+ has grown to 25 million paid subscribers, driven by popular original content such as Ted Lasso and the Oscar-winning film Coda. In addition to paying customers, the platform reaches another 50 million users through free promotional access, highlighting its expanding audience and strong appeal in the competitive streaming market.
One in Four Americans Cancel and Later Resubscribe to Streaming Services
About 25% of Americans “churn and return” to video streaming services, meaning they cancel and later resubscribe. This behavior is even more common among younger generations, with 35% of Gen Z and 38% of Millennials having canceled and re-subscribed to a streaming platform in the past 12 months, showing that switching between services is a frequent trend among younger viewers.
Ad-supported streaming services are growing faster than subscription-based platforms in the U.S. Between 2020 and 2022, households streaming ad-supported video increased by 29%, compared to 22% growth for subscription-based services, according to Comscore.
Platforms like Sling and Tubi offer content without monthly fees, funding their services through ads in a model similar to traditional broadcast TV, making them an increasingly popular option for cost-conscious viewers.
Wrapping Up
The trend of cord cutting is expected to continue growing in the coming years, reshaping the television and entertainment industry. It is expected that most U.S. households with a TV will stop using traditional cable or satellite services.
Streaming platforms will keep growing, adding more shows, movies, and original content, often with flexible pricing to attract viewers. Free options like over-the-air (OTA) TV and ad-supported streaming will also become more popular, offering cheaper alternatives.
Meta released a new social media platform Threads back in July 2023, which created a massive storm in the market. Threads was able to record 1 million users within the first hour of its release and 10 million users in the first 7 hours of its launch. This application was also able to create a milestone of 100 million users in just 4 days and 6 hours.
As of 2024, there are more than 160 million threads users worldwide with 10.3 million daily active users. In this article, we are going to provide an up-to-date guide on Meta’s latest social media platform including the number of threads users (January 2023), Threads user growth, User activity, and much more. So, let’s begin.
Key Statistics of Threads Users
As of 2024, there are over 160 million users worldwide.
Threads crossed 1 million users within one hour of its release.
Threads was able to reach a milestone of 70 million users in just 2 days.
There are around 10.3 million daily active users.
Neymar is the most-followed person on threads with 13.1 million followers.
Selena Gomez is the second most followed person on threads with 12.3 million followers.
The average time spent by a user on threads every day is 3 minutes.
Threads was launched in the European Union on December 14, 2023.
It took 4 days and 6 hours for threads to achieve 100 million users on the platform.
The number of daily active users on threads has dropped by 79%.
Number of Threads users since its launch
Threads was one of the most anticipated social media platforms among the users and its release back in July 2023 resulted in complete chaos when the platform broke the record for fastest platform to reach 1 million users within the first hour of its release. Within the next two hours, Threads recorded 2 hours of users on its platform and the numbers instantly reached 5 million in 4 hours.
The milestone of 10 million users was reached by threads in 7 hours since its launch and after 1 day and 10 hours, the platform recorded 30 million users on Threads. It only took threads 4 days and 6 hours to reach 100 million users on the platform. Another spike in users was witnessed during December as the platform was released in the European Union, where the users were able to sign up on the platform without linking their Instagram accounts.
Below we have mentioned a table showcasing the number of threads users since its launch:
Threads gained instant popularity among users when it was first released back in July 2023. Meta’s latest social media platform was even able to beat several other online services in reaching 1 and 10 million users in the shortest amount of time.
Apparently, Threads created a record of the fastest platform to reach 1 million users within the first hour of its release and reach 10 million users at a whopping speed of 7 hours. Meanwhile, other platforms like ChatGPT took a duration of 5 days to reach 1 million users and about 40 days for 10 million users. Below we have mentioned a table comparing the time taken by threads and other services to reach 1 and 10 million users.
The user activity of threads has witnessed its ups and downs ever since its launch. Initially, threads created a lot of buzz for itself when it came out on July 6. However, within a few months, users’ engagement started to decline. Below we have mentioned all the data regarding the user activity on threads:
Threads gained 41.79 million daily active users on its launch on July 6
Meta Platforms launched Threads worldwide on July 6, closely linked with the popular social media platform Instagram. Within a few hours of its release Threads gained massive popularity on the platform with a large range of users signing up to the platform and it gained 41.79 million daily active users. The numbers witnessed a massive growth on July 7 when it climbed to 49.3 million.
Threads daily active users dropped on July 14
The fall of daily active users on Threads was first reported on July 14 when the number of daily active users dropped by 52% within a week to 23.6 million. Ever since then, the numbers have been significantly decreasing and currently, the platform has around 10.3 million daily active users, which is down by 79% since its peak in July.
The Time spent on Threads has dropped significantly
The average time spent by users on the threads on July 7 was 21 minutes on a daily basis. However, the time dropped significantly ever since, and today the average time spent by users on the platform is 3 minutes.
Meanwhile, the biggest competition of threads Twitter has 200 million daily active users with an average time spent of 31 minutes every day.
Popular Footballer Neymar is currently the most-followed person on Threads where he has 13.1 million followers as of February 2024. Selena Gomez is the second most followed person on Threads with 12.3 million followers. Below we have mentioned a table of the top 10 most followed accounts on Threads:
Rank
Most Followed Accounts
Total Followers
1
Neymar
13.1 million
2
Selena Gomez
12.3 million
3
Kylie Jenner
10.6 million
4
Kim Kardashian
9.4 million
5
Jennifer Lopez
7.5 million
6
Mr. Beast
7.1 million
7
Shakira
6.9 million
8
Will Smith
6.6 million
9
Khloe Kardashian
6.2 million
10
Chris Hemsworth
6 million
FAQs
Do you need an Instagram account for Threads?
Yes, a user must have an Instagram account to create an account on Threads since both platforms are integrated. So, if you want to create an account on threads you must first sign-up on Instagram.
Is Threads app free?
Yes, threads is a free app that can be downloaded by anyone through the App Store or Play Store and can be accessed by users without any cost.
Can you make Threads private?
Yes, threads allow users to set their accounts private. This way only your followers will be able to see your threads or posts and will be permitted to reply to you.
Wrapping Up
The massive popularity of Threads resulted in the application reaching 1 million users within the first hour of its release and 100 million users in 4 days and 6 hours. However, things seem to have slowed down for threads as the number of daily active users dropped by 79%.
Although there was a bit of a rise in December when the application was released in the European Union. But to keep the pace moving threads must introduce unique and engaging features and attributes that can attract a higher audience to the application.
Deepfakes are currently one of the most critical issues we are facing in the world. AI technologies have today made it easier for anyone to create manipulated video, audio and images.
Deepfake tools can swap faces of people in seconds and show events that never happened in real life. Thousands of celebrities, politicians and even ordinary people have faced this situation. The scariest part is that this content can look so realistic that it is difficult to detect.
In such scenarios, AI watermarks can help identify AI-generated content. It can provide information about the origin of a video and help determine if the content was created using AI. However, AI watermarks alone are not enough.
Metadata can be lost when files are edited, converted into another format, or shared across different platforms. Even screenshots can break the link to the original content. Therefore, AI watermarks won’t stop deepfakes unless social media platforms and editing tools themselves can stop breaking them.
Content Credentials and AI Watermarks Have Their Limits
Digital or AI-generated content can be primarily detected through two methods: Content Credentials and Invisible Watermarks. Both methods have different ways to identify digital content.
Content Credentials uses the Coalition for Content Provenance and Authenticity (C2PA) standard. Its advanced authenticity detection tools can record information regarding where the content originally came from and its editing history.
Invisible Watermarks, on the other hand, work differently. This method adds an invisible signal directly into the digital media, such as an image, audio clip or video clip. Based on the technology and the edits made to the file, the invisible signal may still remain detectable when the metadata disappears.
Both Content Credentials and Invisible Watermarks are used in OpenAI for supported images created using its tools. Regardless of using both methods, OpenAI has acknowledged that neither of the methods is 100% reliable.
Metadata is capable of getting lost during file handling or too much editing. This can change or weaken the detectable watermark in the content. Therefore, these methods cannot guarantee a perfect solution for deepfake content detection. Instead, they can be considered a useful method to identify the content but are not completely reliable.
AI watermarks alone are not enough. Platforms must also preserve watermark information when content is uploaded and shared. In order to truly identify deepfakes, the platform needs to build systems that can preserve information that shows where the files came from.
Because when a deepfake image is created, it may have information about the origin or watermarks of the AI generator. However, such information can be easily dropped with each image shared.
Like, a user can download a deepfake image, upload it to different social media platforms, a third person can take a screenshot and share it in different groups. The metadata will disappear with each step. This makes the watermark extremely difficult to detect, which will eventually make it harder to recognize the original source of the image.
This makes it important for platforms to create more effective systems that can keep the metadata protected regardless of the sharing stages.
For example, social media platforms could keep valid Content Credentials when they edit or resize images. They could also check for watermarks and show the results to users. If no watermark is found, the platform should not assume that the content is real or authentic.
However, these steps might not help identify each and every deepfake. But it can still help journalists and fact-checkers identify suspicious content more easily.
How Common Standards Can Improve Deepfake Detection
One of the major issues with AI watermarks and deepfake detection tools is that they do not always use the same platform or follow the same signals. A verification system can successfully detect a watermark on a deepfake image created by a company. However, it may not recognize content created by a completely different AI tool. These limitations can be a huge problem when it comes to identifying deepfake content.
This is why sharing the same standard can help solve this problem. C2PA provides a common way to record where digital content came from and how it has changed. If more companies adopt this standard, different platforms may find it easier to recognise deepfake content.
But just adopting a standard is not enough. Companies need to make sure the systems work properly. They need to focus on what happens to the metadata when files are uploaded to a different platform. Along with the changes when an image is downloaded, edited or converted to another format.
Platforms should focus on these areas, test these processes and explain which signals they can detect. Additionally, they should also provide users with simple methods to check where content came from without any heavy technical knowledge.
The purpose is to make sure information about a deepfake content origin can travel across different platforms. Instead of the information being available only on the platform where the content was originally created.
AI Watermarks Alone Cannot Solve the Deepfake Problem
AI watermarks can be a useful way to identify deepfake content. However, the usefulness depends on the technology itself. AI watermarks should remain intact regardless of the editing and how many times the file is shared to be truly effective. Some content might not have any watermarks at all. AI developers need to work on detection methods to ensure watermarks are detected regardless of which platform was used.
Apart from watermarks and detection tools. Social media and editing platforms should also take responsibility for preserving information about a piece of content’s origin. It’s important to know that not a single method can stop all deepfakes. But AI watermarks and detection tools can certainly help identify manipulated content and take necessary action against.
Right now, the main challenge is not just to add a watermark to an AI-generated file. Instead, it’s to make sure that particular watermark remains in the file after the file has been shared. Until this issue is resolved, watermarks will remain useful but will not be enough to solve the deepfake problem.
Nearly 150 female politicians across Europe have been found listed or depicted on websites featuring sexual deepfakes, according to research reported by WIRED. It covers 22 European Union (EU) countries and highlights the use of artificial intelligence (AI) to create or spread explicit content involving women in public life.
The research identified at least 138 female politicians connected to these websites. Such content can use AI to create fake sexual images or alter real photographs to make them appear explicit. These images can be shared online without the person’s consent.
This raises concerns about online abuse, privacy and the safety of women in politics. They also highlight the challenges of removing non-consensual deepfakes from websites and preventing them from spreading.
Research Finds Female Politicians Across 22 EU Countries
The research identified at least 138 female politicians listed or depicted on sexual deepfake websites across 22 European Union (EU) member states. This suggests that this form of online abuse affects women in public office across several European countries.
Among the politicians publicly identified in reports about the issue are Dutch MPs Suzanne Kröger and Sarah Dobbe. Their cases highlight concerns about the use of politicians’ names, photographs and likenesses on websites featuring explicit content. However, the public findings do not provide a complete list of all the politicians identified in the research.
Sexual deepfakes use AI tools to create or manipulate images and videos with explicit content. In many cases, these tools can place a real person’s face onto a different body or alter existing images. The resulting content may look convincing, even when the person has never taken part in the activity shown.
The presence of politicians on these sites does not necessarily mean that every person identified was depicted in an AI-generated image. Some may have been listed or referenced on the websites. The exact circumstances can differ from one case to another.
However, this shows how websites featuring sexual deepfakes can target women who hold public positions. The content can also reach large audiences, creating additional concerns for the people involved.
European Politicians Targeted by Deepfake Websites
Research by Agora Digitale Transformation found that female politicians across Europe appeared on explicit deepfake websites. Data captured in July 2026 showed differences in the number of members of parliament (MPs) identified in each country.
Germany had the highest number of MPs identified on explicit deepfake websites, with 28, followed by the Netherlands with 22, Italy with 18 and France with 17. Poland recorded 12 MPs, while Finland had nine. Denmark, Sweden, Greece and Belgium recorded seven, six, five and five MPs, respectively.
Countries
MPs Targeted
Germany
28
Netherlands
22
Italy
18
France
17
Poland
12
Finland
9
Denmark
7
Sweden
6
Greece
5
Belgium
5
How Sexual Deepfakes Can Affect Women in Politics
Sexual deepfakes can damage a person’s reputation and expose them to unwanted attention. Even when an image is fake, viewers may believe it is real or share it without checking its authenticity.
For female politicians, this can create additional pressure in their professional and public lives. They may have to respond to false claims, report abusive content or seek legal help to protect their privacy.
The effects can extend beyond the individuals shown in the images. The spread of sexual deepfakes may also discourage women from participating in political discussions or pursuing public office if they fear similar abuse.
However, the impact can vary. Not every targeted person will experience the same consequences, and the research findings alone do not establish the effect on each politician.
Why AI-Generated Sexual Images Are Difficult to Control
AI tools have made it easier to create realistic images with limited technical knowledge. Some tools can alter photographs or generate explicit images from existing pictures. This can allow people to produce misleading content without the consent of the person depicted.
Removing such material can also be difficult. An image may appear on several websites or be reposted across social media platforms. Even after the original content is removed, copies may continue to circulate.
Identifying the person who created or uploaded an image can present further challenges. Websites may operate across national borders, while the people responsible may use anonymous accounts or other methods to hide their identities. These issues can complicate efforts by victims, platforms and authorities to respond to sexual deepfakes.
European Countries Face Questions Over Deepfake Protections
The findings add to wider concerns about how European countries should respond to non-consensual sexual content created or altered using AI.
Possible responses include clearer rules on AI-generated intimate images, faster reporting and removal systems, and stronger support for victims. Law enforcement agencies may also need ways to investigate cases where the people responsible operate across borders.
Technology companies face questions about how they detect and restrict explicit content that uses a real person’s likeness without permission. However, effective enforcement can depend on the laws that apply, the location of the website and the evidence available in each case.
The research does not, by itself, establish that all the websites or images identified broke a particular law. Legal responsibility depends on the circumstances and the rules that apply in each country.
How Sexual Deepfakes Are Raising New Regulatory Concerns
The identification of at least 138 female politicians across 22 EU countries highlights the scale of the issue uncovered by the research. It also shows why discussions about AI safety increasingly include the misuse of image-generation tools.
The challenge extends beyond whether AI can produce realistic images. It also involves consent, the distribution of explicit material and the ability of affected people to seek help.
For policymakers, the findings may add to calls for stronger protections against non-consensual sexual deepfakes. For technology companies, they underline the need to consider how their tools could be misused and what safeguards may reduce that risk.
The research offers evidence of a problem affecting women in public life. Further details about the individual cases, the websites involved and the responses from authorities would help clarify the full extent of the issue.
A U.S. appeals court has temporarily blocked Minnesota’s ban on AI-generated fake nude images. The decision comes as Elon Musk’s xAI challenges the law on constitutional grounds.
The case raises questions about how far states can go to regulate AI-generated sexual images. It also highlights the legal challenges facing governments as they try to prevent the misuse of artificial intelligence.
Minnesota’s law aimed to address the creation and distribution of intimate images made without a person’s consent. Such images can use AI tools to make it appear that a real person is nude, even when the person never posed for the image.
The court’s decision temporarily prevents the ban from taking effect while the legal challenge continues. However, it does not settle the broader debate over AI-generated sexual content or determine the final outcome of the case.
AI nudification tools can alter photographs to create fake nude images of real people. These tools can cause serious harm when someone uses them without the person’s permission.
Victims may face harassment, humiliation, reputational damage and threats. The images can also spread online quickly, making it difficult to remove them completely.
Lawmakers in several U.S. states have considered or introduced measures to address this problem. Their efforts reflect growing concern about AI tools that make it easier to create realistic sexual images without consent.
Minnesota’s ban sought to restrict this type of misuse. However, laws regulating AI-generated content can face legal challenges over free speech and other constitutional protections.
These disputes often depend on how a law defines prohibited content, which activities it restricts and whether its wording meets constitutional standards.
Why xAI Is Challenging Minnesota’s AI Nudification Ban
xAI, the artificial intelligence company founded by Elon Musk, is challenging Minnesota’s law on constitutional grounds.
The dispute raises questions about whether the restrictions place unlawful limits on protected expression. The exact legal arguments and their treatment by the court will be important as the case moves forward.
Laws targeting AI-generated content must balance different interests. Governments want to protect people from non-consensual sexual images. Along with this, restrictions must comply with the U.S. Constitution.
The legal debate may also affect how lawmakers draft future AI regulations. Broad restrictions could face challenges if courts find that they reach protected speech or fail to define prohibited conduct clearly.
However, a constitutional challenge does not automatically mean that every restriction on AI-generated sexual images is unlawful. The outcome depends on the specific law and the court’s legal analysis.
The Impact of the Court’s Ruling on Minnesota’s AI Ban
The appeals court’s decision temporarily blocks Minnesota’s ban while the challenge proceeds. It should not be treated as a final ruling that invalidates the law permanently. The distinction is important. A temporary court order can pause the enforcement of a law without deciding whether the law is constitutional.
The decision also does not mean that creating or sharing non-consensual intimate images is automatically legal in Minnesota. Other state or federal laws may apply, depending on the conduct and circumstances.
For people affected by AI-generated sexual images, the case highlights the uncertainty that can arise when new technology meets existing legal protections. It also shows why the wording of AI laws matters. Legislators must consider how to prevent abuse while ensuring that restrictions can withstand constitutional review.
Implications for State Laws Targeting AI Deepfakes
The outcome could influence how other states approach laws targeting AI-generated intimate images. If the court identifies constitutional problems with Minnesota’s restrictions, lawmakers may need to reconsider how similar measures are written.
Other states may also examine the case when developing rules for AI-generated sexual content. The legal reasoning could help clarify which restrictions are permissible and which may raise First Amendment concerns.
The case is part of a wider debate over responsibility for AI-generated content. Governments are under pressure to respond to deepfakes and other forms of digital abuse. Yet courts must assess whether the measures used to address these problems comply with constitutional protections.
For AI companies, the dispute could also shape expectations around the regulation of tools that generate or alter images. The final ruling may offer guidance on how states can address harmful uses of these systems.
The next key development will be the progress of xAI’s legal challenge and any further decisions from the courts. The final outcome will determine whether Minnesota can enforce its ban in its current form or must revise its approach.
Until then, the temporary block leaves the law’s future uncertain. The case will test how courts balance efforts to prevent non-consensual AI-generated nude images with constitutional protections for speech. For now, the decision is a pause in the legal fight, not a final judgment on the law’s validity.
Italy’s Prime Minister Giorgia Meloni is seeking to trademark her voice in the European Union (EU) to fight the growing threat of AI deepfakes. The move comes after criminals used fake audio and images of her in scams. Her likeness has also appeared in misleading online content.
The proposed trademark could give Meloni another legal tool to challenge the unauthorised use of her voice. It also highlights growing concerns about artificial intelligence (AI) tools that can copy a person’s voice and create realistic fake recordings.
As AI-generated content becomes harder to distinguish from genuine recordings, public figures face new risks. Fake audio can make it appear that a person said something they never said. Such content can also be used to spread false information or trick people into sending money.
Why Giorgia Meloni Wants to Trademark Her Voice
Meloni’s move follows cases in which her image and voice were used to create misleading content. AI tools can reproduce a person’s speech, tone and manner of speaking with only a limited amount of audio. This makes it easier for criminals to create convincing fake recordings.
Voice cloning can also make scams more difficult to detect. A fake recording may appear to come from a trusted public figure, business leader or family member. People who believe the recording is genuine may be more likely to trust its message.
For politicians, the risks extend beyond financial fraud. Fake audio can also spread false statements, damage reputations and influence public opinion. A realistic recording may circulate online before fact-checkers or the person involved can respond.
By seeking trademark protection for her voice, Meloni is looking for another way to challenge its unauthorised use. However, the legal effect will depend on the scope of the application and the protection granted under applicable law.
How AI Voice Cloning and Deepfakes Are Used for Scams
AI deepfakes use artificial intelligence to create or alter audio, images and videos. Some tools can imitate a person’s voice from existing recordings. Others can generate videos that make someone appear to say or do things that never happened.
Scammers can use these tools to create false endorsements, fake public statements or fraudulent messages. A cloned voice may be used to promote an investment scheme or persuade people to transfer money.
Political figures are also vulnerable to this type of misuse. Fake recordings can falsely link them to statements, decisions or events. Even after a recording is exposed as fake, it may continue to circulate on social media.
The problem is not limited to well-known people. Criminals can also clone the voices of ordinary individuals. This creates risks for families, businesses and anyone whose voice recordings are publicly available.
Trademark protection generally helps distinguish the goods or services of one person or business from those of others. The EU has a trademark registration system managed by the European Union Intellectual Property Office (EUIPO).
However, registering a voice is not the same as gaining complete ownership of every recording that sounds similar. The legal protection depends on the type of trademark, the details of the application and the rules that apply.
A trademark may help its owner challenge certain unauthorised commercial uses. But it does not automatically make every AI-generated imitation illegal. Other legal rules may apply to fraud, privacy, copyright, image rights and the misuse of personal data.
The outcome of Meloni’s application will therefore matter. Its scope and any conditions attached to registration will help determine how useful it could be against AI-generated imitations.
What Meloni’s Move Means for AI Deepfake Regulation
Meloni’s effort reflects a wider debate over how existing laws should address AI-generated content. Governments and regulators are under pressure to protect people from impersonation while preserving legitimate uses of the technology.
Deepfakes can serve lawful purposes in films, satire, entertainment and political commentary. The challenge is to distinguish these uses from content created to deceive people, damage reputations or commit fraud.
Trademark protection could offer public figures another legal route to challenge certain forms of misuse. However, it is unlikely to solve the wider deepfake problem on its own.
Platforms may still need effective systems to identify misleading content and respond to complaints. People also need ways to verify suspicious recordings before trusting or sharing them.
Meloni’s Trademark Bid Puts the Focus on the Future of Voice Protection
The next key development is the outcome of Meloni’s trademark application in the EU. The details will determine whether the registration offers meaningful protection against unauthorised uses of her voice.
Even if the application succeeds, enforcement may remain difficult. AI-generated recordings can spread across multiple platforms and countries within minutes. Identifying the people responsible can also be challenging.
Meloni’s move highlights a growing legal challenge as AI tools make voice imitation easier. It could add to the debate over whether existing intellectual property rules are sufficient to protect people from deepfakes.
For now, the case serves as a reminder that a familiar voice is no longer proof that a recording is genuine. Verifying the source of suspicious audio is becoming increasingly important as voice-cloning technology advances.
AI-generated deepfakes are creating a new legal challenge for U.S. election candidates. Political campaigns are increasingly facing disputes over manipulated images, videos and audio recordings. Some candidates are responding with legal notices, criminal complaints and lawsuits.
These cases show how AI deepfakes are moving beyond online misinformation. They are becoming a legal issue for candidates, political parties and the people who create or share misleading content.
The growing use of AI tools has made it easier to imitate a politician’s face or voice. A fake recording can make a candidate appear to say something they never said. A manipulated video can also spread false claims about a campaign.
As these incidents become more common, candidates are looking for legal ways to respond. Their actions could also shape how courts handle AI-generated political content in future elections.
U.S. Politicians Turn to Lawsuits to Fight AI Deepfakes
Political candidates in the United States are increasingly turning to legal action over manipulated campaign content. Responses can include cease-and-desist letters, criminal complaints and civil lawsuits.
A cease-and-desist letter asks a person or organisation to stop a specific activity. In a deepfake dispute, it may demand that the recipient remove misleading content or stop distributing it. However, receiving such a letter does not automatically mean that a court has found the content illegal.
Criminal complaints follow a different process. They ask law enforcement authorities to investigate suspected wrongdoing. Prosecutors must then determine whether the evidence supports criminal charges under applicable law.
Civil lawsuits can give candidates another way to challenge harmful content. Depending on the facts, a case may involve claims such as defamation or other violations of state law.
However, legal action does not guarantee that a candidate will win. The outcome depends on the content, the evidence, the relevant laws and the circumstances in which the material was created or shared.
How AI Deepfakes Can Affect Election Campaigns
AI deepfakes can make false political messages appear convincing. Tools can generate realistic voices, alter facial expressions or create videos that show events that never happened.
For example, a fake audio clip could make a candidate appear to endorse a controversial policy. A manipulated video could also make it seem as if a politician made an offensive statement. Such content can spread quickly through social media and messaging platforms. Voters may encounter a clip without seeing the original source or a warning that it has been altered.
The timing can make the problem harder to address. A misleading video released shortly before an election may circulate widely before a campaign can respond. Even if the content is later removed or disproved, some voters may continue to believe it.
Deepfakes can also create confusion around genuine political footage. When voters become less certain about what they see and hear, authentic recordings may face greater suspicion. Still, not every AI-generated political image or video is illegal.
Satire, parody and other forms of political expression can receive legal protection. The key questions include whether the content is misleading, how it is used and whether it violates an applicable law.
Taking legal action against an AI deepfake can be challenging. Candidates must identify the people responsible, gather evidence and establish a legal basis for their claims.
Defamation cases can be particularly difficult. Public figures generally face a high legal standard when suing over statements about them. In many cases, they must show that the defendant acted with actual malice. This means knowing that a statement was false or acting with reckless disregard for whether it was true.
However, the standard depends on the claim and the circumstances. A manipulated video may raise different legal questions from a false written statement or an unauthorised use of someone’s voice.
Identifying the creator can also be difficult. A deepfake may be shared through anonymous accounts or reposted by multiple users. The person who created the content may not be the same person who distributed it widely.
There is also a question of jurisdiction. U.S. states have different laws governing elections, impersonation and deceptive media. A case may depend on where the content was produced, how it was distributed and which laws apply.
These challenges can make legal action slow and expensive. They can also limit how quickly a candidate can respond to a deepfake during an active campaign.
Recent Cases Show How Deepfake Disputes Are Escalating
Several disputes during the 2026 U.S. election cycle illustrate how candidates are responding to AI-generated campaign content.
Derrick Van Orden Faces a Cease-and-Desist Letter
Attorneys for Rebecca Cooke, a Democratic candidate challenging Republican Representative Derrick Van Orden in Wisconsin, sent him a cease-and-desist letter in October 2026. They objected to AI-generated videos that allegedly misrepresented Cooke’s political views.
One video showed an AI-generated version of Cooke making statements that her campaign said were false. Van Orden responded publicly with another AI-generated video, this time featuring a cartoon duck. The dispute highlights the difficulty of separating political satire from misleading content during an election campaign.
Dan Patrick Faces a Criminal Complaint in Texas
Texas Lieutenant Governor Dan Patrick also faced a criminal complaint over AI-generated political advertisements. His Democratic opponent, state Representative Vikki Goodwin, filed the complaint under Texas’ deepfake election law.
The advertisements allegedly used manipulated footage to depict Goodwin at a football game wearing a shirt that said “taxes” instead of “Texas.” Patrick’s campaign later removed the ads. A spokesperson described the video as an obvious parody.
The complaint raises questions about how election laws apply to AI-generated attack ads. It also shows how political opponents can use legal procedures to challenge content they believe crosses the line.
These cases involve allegations and legal complaints. They should not be treated as proof that a court has found the accused parties guilty of wrongdoing.
U.S. Election Deepfake Laws Remain Uneven
The United States does not have a single, comprehensive federal standard governing all AI-generated political campaign content. Instead, candidates and campaigns must consider a mix of state laws and existing legal protections.
According to Axios, about 31 states regulate political deepfakes. At least 28 require disclosures for AI-generated content in certain circumstances.
These rules vary. Some focus on realistic content that could mislead voters close to an election. Others require political advertisements to disclose when they use AI-generated or manipulated material.
The differences can make enforcement complicated. A campaign tactic that violates one state’s law may face different restrictions in another state.
Free speech protections also play an important role. Political satire and criticism are central parts of U.S. elections. Broad restrictions on manipulated content could raise constitutional concerns if they also limit protected expression.
Lawmakers therefore face a difficult task. They must consider ways to protect voters from deliberate deception without preventing legitimate political commentary or parody.
How AI Deepfake Lawsuits Could Change Future U.S. Elections
The growing number of complaints and legal threats suggests that candidates are becoming less willing to ignore misleading AI content. Campaigns may increasingly rely on lawyers, election officials and courts to challenge material they believe breaks the law.
These disputes could also influence how political advertisements are produced. Campaigns and outside groups may need to pay closer attention to disclosure requirements, state election laws and the risks of publishing realistic but misleading videos.
However, legal action may not stop a deepfake from spreading immediately. Court proceedings can take time, while manipulated content can reach thousands of people within minutes. For now, the 2026 election cycle is showing that AI deepfakes are becoming more than a problem for campaign communications teams. They are also creating legal battles over deception, political expression and accountability.
As more candidates challenge manipulated campaign material, their cases could help define the limits of AI use in American elections.
