Meta is exploring the possibility of entering paid agreements with news publishers to access content for training its generative AI tools. Internal discussions within Meta involve the partnerships, product and legal teams, as the company seeks to improve its AI models. This move could mark a significant shift from Meta’s recent retreat from the news industry, where it slashed a $2 billion news budget. Despite Meta CEO Mark Zuckerberg’s claim that the company has sufficient training data, concerns persist about its quality. User-generated content from Facebook and Instagram may not provide the high-quality data necessary for advanced AI models. In contrast, content such as news articles, books, and essays is crucial for generating accurate outputs. Meta’s competitors, including Google and OpenAI, have secured deals with major news publishers, positioning themselves ahead in the AI race. Meta has yet to approach any news outlets formally, but the potential for licensing agreements is on the table. As the US Copyright Office considers new rules for generative AI, the pressure mounts for Meta to secure reliable training data.
![meta-considering-deals-with-news-publishers-for-ai-training-data](https://traderstarter.com/wp-content/uploads/2024/05/21770-meta-considering-deals-with-news-publishers-for-ai-training-data.jpg)