Close Menu
TechBrunchTechBrunch
  • Home
  • AI
  • Apps
  • Crypto
  • Security
  • Startups
  • TechCrunch
  • Venture

Subscribe to Updates

Subscribe to our newsletter and never miss our latest news

Subscribe my Newsletter for New Posts & tips Let's stay updated!

What's Hot

Spotify says iOS external payment support is already boosting subscriptions

May 21, 2025

Wyden: AT&T, T-Mobile and Verizon did not inform senators of surveillance requests

May 21, 2025

Spotify's new “future release” hub highlights future albums that you might like

May 21, 2025
Facebook X (Twitter) Instagram
TechBrunchTechBrunch
  • Home
  • AI

    OpenAI seeks to extend human lifespans with the help of longevity startups

    January 17, 2025

    Farewell to the $200 million woolly mammoth and TikTok

    January 17, 2025

    Nord Security founder launches Nexos.ai to help enterprises move AI projects from pilot to production

    January 17, 2025

    Data proves it remains difficult for startups to raise capital, even though VCs invested $75 billion in the fourth quarter

    January 16, 2025

    Apple suspends AI notification summaries for news after generating false alerts

    January 16, 2025
  • Apps

    Spotify says iOS external payment support is already boosting subscriptions

    May 21, 2025

    Spotify's new “future release” hub highlights future albums that you might like

    May 21, 2025

    Ted's app launches short video features ahead of Tiktok's potential ban

    May 21, 2025

    siro lands $50 million to expand AI-powered coaching for salespeople

    May 21, 2025

    Apollo for Reddit Dev Christian Selig joins Digg as an advisor

    May 21, 2025
  • Crypto

    Early savings for 2025 will end on May 25th

    May 21, 2025

    Coinbase says its data breach will affect at least 69,000 customers

    May 21, 2025

    There are 6 days to save $900 to destroy 2025 tickets

    May 20, 2025

    Save $900 to destroy 2025 tickets before prices rise on May 25th

    May 19, 2025

    Crypto Elite is increasingly worried about their personal safety

    May 18, 2025
  • Security

    Wyden: AT&T, T-Mobile and Verizon did not inform senators of surveillance requests

    May 21, 2025

    US students agree to plead guilty to hacking affecting tens of millions of students

    May 21, 2025

    The people in Elon Musk’s DOGE universe

    May 20, 2025

    Cocospy Stalkerware App goes offline after a data breach

    May 19, 2025

    Pharma Giant Regeneron buys 23andMe and its customers' data for $256 million

    May 19, 2025
  • Startups

    7 days left: Founders and VCs save over $300 on all stage passes

    March 24, 2025

    AI chip startup Furiosaai reportedly rejecting $800 million acquisition offer from Meta

    March 24, 2025

    20 Hottest Open Source Startups of 2024

    March 22, 2025

    Andrill may build a weapons factory in the UK

    March 21, 2025

    Startup Weekly: Wiz bets paid off at M&A Rich Week

    March 21, 2025
  • TechCrunch

    OpenSea takes a long-term view with a focus on UX despite NFT sales remaining low

    February 8, 2024

    AI will save software companies' growth dreams

    February 8, 2024

    B2B and B2C are not about who buys, but how you sell

    February 5, 2024

    It's time for venture capital to break away from fast fashion

    February 3, 2024

    a16z's Chris Dixon believes it's time to focus on blockchain use cases rather than speculation

    February 2, 2024
  • Venture

    Submitted submission raises $17 million to automate tax preparation dr voyages

    May 21, 2025

    In a busy VC landscape, Elizabeth Weil's graffiti venture shows that networks are still important

    May 21, 2025

    A comprehensive list of 2025 tech layoffs

    May 21, 2025

    Taylored Side Events will be held at all stages 2025 in Boston

    May 20, 2025

    Adaptation Ventures is a new Angel Investor Group focusing on obstacles and accessibility technologies

    May 20, 2025
TechBrunchTechBrunch

Meta's New AI Deepfake Playbook: More Labels, Less Removals

TechBrunchBy TechBrunchApril 5, 20246 Mins Read
Facebook Twitter Pinterest Telegram LinkedIn Tumblr WhatsApp Email
Share
Facebook Twitter LinkedIn Pinterest Telegram Email


Meta has announced changes to its rules regarding AI-generated content and manipulated media following criticism from the Oversight Board. Starting next month, the company announced it will label a wider range of content, including applying the “Made with AI” badge to deepfakes. Additional contextual information may be displayed if the content is manipulated in other ways that pose a high risk of misleading the public about important issues.

The move could lead to the social networking giant adding more labels to potentially misleading content, which is important in a year with many elections around the world. However, in the case of deepfakes, Meta will only label it if the content in question has an “industry standard AI image indicator” or if the uploader is publishing AI-generated content. Apply.

These out-of-scope content generated by AI will likely be leaked without labels.

This policy change will also likely lead to more AI-generated content and manipulated media remaining on Meta's platform. Meta is shifting toward favoring an approach that focuses on “transparency and providing additional context” as a “better way to deal with this content” (as it reduces the associated risks to free speech). (instead of deleting the manipulated media).

So the strategy for AI-generated or otherwise manipulated media on meta-platforms like Facebook and Instagram seems to be more labels and less removals.

Meta said it would stop removing content based solely on its current manipulated video policy in July, adding in a blog post published Friday: Manipulated media. ”

This change in approach may be aimed at responding to growing legal demands on meta around content moderation and systemic risk, such as the European Union's Digital Services Act. Since August last year, EU law has applied a series of rules to the EU's two major social networks, with Meta having to walk a fine line between removing illegal content, reducing system risks and protecting free speech. I am forced to make a decision. The European Union is also putting extra pressure on platforms ahead of European Parliament elections this June, including asking tech giants to watermark their deepfakes where technically possible. .

The US presidential election scheduled for November is also likely on Mehta's mind.

Criticism of the monitoring committee

Meta's advisory board, which is funded by the tech giant but is allowed to operate independently, reviews a small number of content moderation decisions but makes policy recommendations. You can also do Although Meta is not obligated to accept the board's recommendations, it has agreed to modify its approach in this case.

Monika Bickert, Meta's vice president of content policy, said in a blog post published Friday that the company will revise its policies regarding AI-generated content and manipulated media based on board feedback. “Our existing approach is too narrow because it only targets videos that have been created or modified by AI to appear as if the human is saying something that is not said. “We agree with the committee’s arguments,” she wrote.

Back in February, the Oversight Committee asked Meta to change its approach to AI-generated content in the wake of the doctored video of President Biden, in which a platonic kiss with his granddaughter was edited to suggest a sexual motive. I urged them to reconsider.

The board agreed with Mehta's decision not to publish certain content, but attacked the company's manipulated media policy as “incoherent.” For example, it pointed out that the policy only applies to videos created by AI, while other fake content (such as more basic ones) is allowed. (processed video or audio) off the hook.

Mehta appears to have taken the critical feedback on board.

“Over the past four years, and especially in the last year, this technology has rapidly evolved with the development of other types of realistic AI-generated content, such as audio and photos,” Bickert wrote. “As the Board noted, it is equally important to address manipulation that makes it appear that people are doing something they are not doing.

“The Board also argued that when removing manipulated media that does not violate community standards, it risks unnecessarily restricting freedom of expression. It recommended a “less restrictive” approach. ”

Earlier this year, Meta announced that it was working with others in the industry to develop common technical standards for identifying AI content, including video and audio. The company is now relying on this effort to expand labeling of synthetic media.

“Our 'Made with AI' label for AI-generated video, audio, and images is based on detecting industry-shared signals of people who self-disclose that they are uploading AI images or AI-generated content. ,” Bickert said. Note that the company is already applying the “Imagined with AI” label to photorealistic images created using its proprietary Meta AI capabilities.

Bickert said the expanded policy covers “a broader range of content in addition to the manipulated content that the Oversight Committee recommended labeling.”

“If we determine that digitally created or altered images, video, or audio pose a particularly high risk of materially misleading the public about material matters, we will provide the public with more information and context. may add more prominent labels to the list,” she wrote. “This holistic approach allows people to be more informed about the content, so they can evaluate it better and gain context when they see the same content elsewhere. It will be.”

Meta says it will not remove manipulated content (whether AI-based or otherwise defaced) unless it violates other policies (such as voter interference, bullying or harassment, violence or incitement, or other community standards issues). Ta. Instead, as mentioned above, we may add “informational labels and context” in certain scenarios of high public interest.

Meta's blog post highlights the network of nearly 100 independent fact checkers the company says it engages to help identify risks related to manipulated content.

Per Meta, these external entities will continue to review false and misleading content generated by AI. If content is rated as “false or altered,” Meta said it will respond by applying algorithm changes that reduce the content's reach. This means fewer people see your content because it appears at the bottom of the feed, and the meta also places an overlay label with additional information. The eyeball that lands there.

As synthetic content proliferates due to the boom in generative AI tools, these third-party fact-checkers are likely to face an increased burden. And because we think a lot of this stuff will remain on Meta's platform as a result of this policy change.



Source link

Share. Facebook Twitter Pinterest LinkedIn Tumblr Email

Related Posts

OpenAI seeks to extend human lifespans with the help of longevity startups

January 17, 2025

Farewell to the $200 million woolly mammoth and TikTok

January 17, 2025

Nord Security founder launches Nexos.ai to help enterprises move AI projects from pilot to production

January 17, 2025

Data proves it remains difficult for startups to raise capital, even though VCs invested $75 billion in the fourth quarter

January 16, 2025

Apple suspends AI notification summaries for news after generating false alerts

January 16, 2025

Nvidia releases more tools and guardrails to help enterprises adopt AI agents

January 16, 2025

Leave A Reply Cancel Reply

Top Reviews
Editors Picks

7 days left: Founders and VCs save over $300 on all stage passes

March 24, 2025

AI chip startup Furiosaai reportedly rejecting $800 million acquisition offer from Meta

March 24, 2025

20 Hottest Open Source Startups of 2024

March 22, 2025

Andrill may build a weapons factory in the UK

March 21, 2025
About Us
About Us

Welcome to Tech Brunch, your go-to destination for cutting-edge insights, news, and analysis in the fields of Artificial Intelligence (AI), Cryptocurrency, Technology, and Startups. At Tech Brunch, we are passionate about exploring the latest trends, innovations, and developments shaping the future of these dynamic industries.

Our Picks

Spotify says iOS external payment support is already boosting subscriptions

May 21, 2025

Wyden: AT&T, T-Mobile and Verizon did not inform senators of surveillance requests

May 21, 2025

Spotify's new “future release” hub highlights future albums that you might like

May 21, 2025

Subscribe to Updates

Subscribe to our newsletter and never miss our latest news

Subscribe my Newsletter for New Posts & tips Let's stay updated!

© 2025 TechBrunch. Designed by TechBrunch.
  • Home
  • About Tech Brunch
  • Advertise with Tech Brunch
  • Contact us
  • DMCA Notice
  • Privacy Policy
  • Terms of Use

Type above and press Enter to search. Press Esc to cancel.