Facebook posts from as far back as 2007 being used to train Meta AI
Text and photos posted publicly to Facebook and Instagram from as far back as 2007 are being used to train Meta's artificial intelligence models, a Parliamentary committee has heard.
Representatives from Meta, the owner of Facebook and Instagram, were pressed on their AI models during a Senate select committee hearing on Wednesday. Melinda Claybaugh, global privacy policy director at Meta, confirmed that unless a person had set their profile or posts to private, their social media posts were being used to train Meta's AI models.
Privacy Concerns
Greens Senator David Shoebridge raised concerns about the use of public posts without explicit consent. He highlighted that Meta was scraping all public posts on Instagram and Facebook dating back to 2007, unless specifically set to private.
Senator Shoebridge emphasized the ethical implications, questioning the use of personal data, including images and text, without individuals' awareness or permission.
Ms. Claybaugh clarified that Meta does not utilize data from accounts of individuals under 18 years of age for training its AI models.
Ethical Considerations
Senator Shoebridge further challenged the ethical aspects of using personal posts for AI training, citing scenarios where individuals' moments of celebration and family milestones were unwittingly incorporated into Meta's models.
He urged Meta to address the ethical concerns surrounding the use of user-generated content without explicit consent and the potential invasion of privacy.
In response, Ms. Claybaugh outlined Meta's privacy protocols to prevent the misuse of personal data in AI applications.
Opt-In Process
Labor Senator Varun Ghosh proposed the idea of implementing an opt-in process for users to consent to the use of their data for AI training purposes. However, Simon Milner, vice president of public policy APAC at Meta, expressed reservations about compulsory opt-ins, citing potential user dissatisfaction.
Copyrighted Works
The committee also discussed the use of copyrighted material, including pirated books, to train Meta's AI models. Concerns were raised about the legality of using copyrighted works in AI training processes, with specific references to Australian authors' works.
Meta representatives refrained from commenting on the issue due to its legal complexities in the United States.
Senator Sheldon criticized Meta's data usage practices, labeling them as "dishonest" and "predatory," particularly in the context of utilizing personal information for AI training without explicit consent.
He called for a reevaluation of privacy laws to address the misuse of personal data in AI development.