More than 600,000 people, including several celebrities, have fallen for an online hoax by reposting a fake Instagram message that claims to deny Meta the right to use their images for AI training purposes.

NFL star Tom Brady, actor James McAvoy and actress Julianne Moore are among those who shared the “Goodbye Meta AI” viral story template, as they through that by posting the message they would stop their data from being used to train AI.

“Goodbye Meta AI. Please note an attorney has advised us to put this on, failure to do so may result in legal consequences,” the text states.

“As Meta is now a public entity all members must post a similar statement. If you do not post at least once it will be assumed you are okay with them using your information and photos. I do not give Meta or anyone else permission to use any of my personal data, profile information or photos.”

Instagram labelled the post as “false information”, and said: “The same information was reviewed by independent fact-checkers in another post”. The company also confirmed that posting the message has no impact on any user’s privacy settings.

“Sharing this story does not count as a valid form of objection,” a Meta spokesperson told the BBC.

Lead Stories, one of the Meta’s third-party fact checkers, confirmed that these messages contained ‘false information’. It added that the post originated on 1 September, with GoogleTrends reporting deep surge in searches for the phrase “Goodbye Meta AI” after 24 September. Facebook and Instagram users who don’t want to share their data for AI training models can change their account settings instead.

In September, Meta announced that it started to use public posts to train its AI model after negotiations with British authorities.

“We will begin training for AI at Meta using public content shared by adults on Facebook and Instagram in the UK over the coming months. This means that our generative AI models will reflect British culture, history, and idiom, and that UK companies and institutions will be able to utilise the latest technology.”

In order to build AI models that can understand context, and produce accurate responses, Meta, and every other AI development company, needs human interaction as input, so that the system can develop an understanding of how people actually talk to each other, and refine its outputs based on such.

Concerns for the use of private content for AI training purposes have been recently growing, with Microsoft-owned LinkedIn pausing its training generative AI models using data from UK users earlier this month.

The move came after data protection regulatory concerns were raised by the Information Commissioner’s Office (ICO).

The ICO confirmed LinkedIn had “suspended” AI model training based on data from UK users on Friday, adding it was “pleased” with LinkedIn’s move.

“We are pleased that LinkedIn has reflected on the concerns we raised about its approach to training generative AI models with information relating to its UK users,” said the ICO’s Executive Director, Stephen Almond.

---