The tech giant is facing potential litigation over the use of copyrighted materials in training its AI models. Clegg expects legal disputes over whether creative content is covered by existing fair use doctrine. In response to concerns about the reproduction of copyrighted imagery, Meta has updated its terms of service to prohibit users from generating content that violates privacy and intellectual property rights.
Key takeaways:
- Meta Platforms used public Facebook and Instagram posts to train its new Meta AI virtual assistant, but excluded private posts and chats to respect users' privacy, according to Meta President of Global Affairs Nick Clegg.
- The company took steps to filter private details from public datasets used for training, and did not use content from sites like LinkedIn due to privacy concerns.
- Meta AI, the most significant product among the company's first consumer-facing AI tools, was made using a custom model based on the Llama 2 large language model and a new model called Emu that generates images in response to text prompts.
- Meta has imposed safety restrictions on what content the Meta AI tool could generate, like a ban on the creation of photo-realistic images of public figures, and expects litigation over the use of copyrighted materials in AI training.