logo
Facebook is starting to feed its AI with private, unpublished photos

Facebook is starting to feed its AI with private, unpublished photos

Ammona day ago

Ammon News - For years, Meta trained its AI programs using the billions of public images uploaded by users onto Facebook and Instagram's servers. Now, it's also hoping to access the billions of images that users haven't uploaded to those servers. Meta tells The Verge that it's not currently training its AI models on those photos, but it would not answer our questions about whether it might do so in future, or what rights it will hold over your camera roll images.
On Friday, TechCrunch reported that Facebook users trying to post something on the Story feature have encountered pop-up messages asking if they'd like to opt into 'cloud processing', which would allow Facebook to 'select media from your camera roll and upload it to our cloud on a regular basis', to generate 'ideas like collages, recaps, AI restyling or themes like birthdays or graduations.'
By allowing this feature, the message continues, users are agreeing to Meta AI terms, which allows their AI to analyze 'media and facial features' of those unpublished photos, as well as the date said photos were taken, and the presence of other people or objects in them. You further grant Meta the right to 'retain and use' that personal information.
Meta recently acknowledged that it scraped the data from all the content that's been published on Facebook and Instagram since 2007 to train its generative AI models. Though the company stated that it's only used public posts uploaded from adult users over the age of 18, it has long been vague about exactly what 'public' entails, as well as what counted as an 'adult user' in 2007. The Verge

Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Meta reportedly hires four more researchers from OpenAI
Meta reportedly hires four more researchers from OpenAI

Ammon

timea day ago

  • Ammon

Meta reportedly hires four more researchers from OpenAI

Ammon News - Looks like Meta isn't done poaching talent from OpenAI. Earlier this week, TechCrunch reported that Meta had hired influential OpenAI researcher Trapit Bansal, and according to The Wall Street Journal, it also hired three other researchers from the company. Now The Information is reporting four more Meta hires from OpenAI: Researchers Shengjia Zhao, Jiahui Yu, Shuchao Bi, and Hongyu Ren. This hiring spree comes after the April launch of Meta's Llama 4 AI models, which reportedly did not perform as well as CEO Mark Zuckerberg had hoped. (The company was also criticized over the version of Llama that it used for a popular benchmark.) There's been some back-and-forth between the two companies, with OpenAI CEO Sam Altman suggesting that Meta was offering '$100 million signing bonuses' while adding that 'so far, none of our best people' have left. Meta CTO Andrew Bosworth then told employees that while senior leaders may have been offered that kind of money, 'the actual terms of the offer' were more complex than a simple one-time signing bonus.

Facebook is starting to feed its AI with private, unpublished photos
Facebook is starting to feed its AI with private, unpublished photos

Ammon

timea day ago

  • Ammon

Facebook is starting to feed its AI with private, unpublished photos

Ammon News - For years, Meta trained its AI programs using the billions of public images uploaded by users onto Facebook and Instagram's servers. Now, it's also hoping to access the billions of images that users haven't uploaded to those servers. Meta tells The Verge that it's not currently training its AI models on those photos, but it would not answer our questions about whether it might do so in future, or what rights it will hold over your camera roll images. On Friday, TechCrunch reported that Facebook users trying to post something on the Story feature have encountered pop-up messages asking if they'd like to opt into 'cloud processing', which would allow Facebook to 'select media from your camera roll and upload it to our cloud on a regular basis', to generate 'ideas like collages, recaps, AI restyling or themes like birthdays or graduations.' By allowing this feature, the message continues, users are agreeing to Meta AI terms, which allows their AI to analyze 'media and facial features' of those unpublished photos, as well as the date said photos were taken, and the presence of other people or objects in them. You further grant Meta the right to 'retain and use' that personal information. Meta recently acknowledged that it scraped the data from all the content that's been published on Facebook and Instagram since 2007 to train its generative AI models. Though the company stated that it's only used public posts uploaded from adult users over the age of 18, it has long been vague about exactly what 'public' entails, as well as what counted as an 'adult user' in 2007. The Verge

Meta fends off authors' US copyright lawsuit over AI
Meta fends off authors' US copyright lawsuit over AI

Ammon

time4 days ago

  • Ammon

Meta fends off authors' US copyright lawsuit over AI

Ammon News - A federal judge ruled on Wednesday for Meta Platforms against a group of authors who had argued that its use of their books without permission to train its artificial intelligence system infringed their copyrights. U.S. District Judge Vince Chhabria, in San Francisco, said in his decision, opens new tab that the authors had not presented enough evidence that Meta's AI would dilute the market for their work to show that the company's conduct was illegal under U.S. copyright law. Chhabria also said, however, that using copyrighted work without permission to train AI would be unlawful in "many circumstances," splitting with another federal judge in San Francisco who found on Monday in a separate lawsuit that Anthropic's AI training made "fair use" of copyrighted materials. "This ruling does not stand for the proposition that Meta's use of copyrighted materials to train its language models is lawful," Chhabria said. "It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one." A spokesperson for the authors' law firm Boies Schiller Flexner said that it disagreed with the judge's decision to rule for Meta despite the "undisputed record" of the company's "historically unprecedented pirating of copyrighted works." A Meta spokesperson said the company appreciated the decision and called fair use a "vital legal framework" for building "transformative" AI technology. The authors sued Meta in 2023, arguing the company misused pirated versions of their books to train its AI system Llama without permission or compensation.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store