Meta recently upgraded its help resource centre with a form that gives users some say in what personal details are employed to teach generative artificial intelligence systems. Nevertheless, the form does not take into account user data from Facebook properties like comments on Facebook and Instagram photos. Recently, a consortium of global data protection agencies issued a collective statement concerning data scraping and safeguarding people’s privacy to organisations like Meta, Alphabet, and Microsoft. enforcement
Facebook users are now able to request that some of their personal data which is used by the company for training general artificial intelligence models be accessed, altered, or deleted. This week, Meta updated its Facebook help center resource section with a form titled "Generative AI Data Subject Rights" to allow people to submit such requests. Generative AI technology is becoming more widespread across the tech industry, with companies using it to create more advanced chatbots and transform simple text into complex answers and images. With this opt-out tool, Facebook users can now choose to access, change, or delete personal data from the third-party sources the company uses to develop its large language and AI models.
Meta is using data which is "publicly available on the internet or licensed sources" to train its generative AI models. This includes blog posts, which could contain personal information such as names and contact information. This form does not cover activity on Meta-owned properties like Facebook comments or Instagram posts; however, Meta is still able to potentially use such first-party data to train its generative AI models.
The company is clear that their Llama 2 open-source large language model was not trained on Meta user data, and that they have not launched any Generative AI consumer features yet. Meta has specified that depending on where a person is located, they may be able to use their data subject rights to opt-out of having certain data used to train their AI models, due to various data privacy regulations outside of the United States.
Meta is like many other tech firms in that it gathers enormous amounts of third-party data to train its AI models. To do this, they are taking public information from the web, as well as securing data from other providers. Although they assert that their use of public and licensed data is in their interest and that they are committed to being transparent regarding the legal basis they have for processing the data, some data privacy advocates have raised concerns about collecting vast quantities of publicly available information for AI model training.
Recently, a consortium of data protection agencies from the U.K., Canada, Switzerland and other countries wrote a joint statement to Meta, Alphabet, TikTok parent ByteDance, X (formerly known as Twitter), Microsoft and others about data scraping and protecting user privacy to remind them of the applicable data protection and privacy laws in different countries.
To delete some of the Facebook data used for generative AI training, users can go to the Generative AI Data Subject Rights form on Meta's privacy policy page, click the link to 'Learn more and submit requests here', and select one of three options that best describes their issue/objection. After selecting an option, users must pass a security check test. Some users have commented that they are encountering issues with completing the form due to what appears to be a software bug.
コメント