Meta has disclosed that its new AI virtual assistant, Meta AI, was developed using public posts from its users on Facebook and Instagram.
Despite this, Nick Clegg, Meta’s President of Global Affairs, emphasized that they did not use private chats from Facebook Messenger or Instagram, respecting user privacy.
Clegg indicated that most of the data Meta used for AI training was publicly accessible, and datasets containing a significant amount of personal information were intentionally left out.
This statement aims to address concerns about tech giants such as Meta, OpenAI, and Alphabet using online data without explicit consent to enhance their AI systems.
Introduced by Mark Zuckerberg during Meta’s annual product conference, Connect, Meta AI is grounded on the Llama 2 large language model and a new image-generating model named Emu.
Clegg revealed that this AI product would produce text, audio, and images, and will be integrated with Microsoft’s Bing search engine for real-time data.
He also mentioned that both text and images from public Facebook and Instagram posts contributed to training Meta AI.
Lastly, Clegg confirmed safety measures in place to regulate the AI’s content generation, including prohibiting it from crafting photo-realistic images of public personalities.
Speaking on copyrighted materials, Clegg said he was expecting a “fair amount of litigation” over the matter of “whether creative content is covered or not by existing fair use doctrine,” which permits the limited use of protected works for purposes such as commentary, research, and parody.
“We think it is, but I strongly suspect that’s going to play out in litigation,” Clegg noted.
