Meta Eyes Paid Deals With News Publishers for AI Training
Meta is exploring paid deals with news publishers for AI training to enhance its generative AI tools. This strategic move follows OpenAI’s recent partnership with News Corp, highlighting the competitive AI landscape.
- Meta Platforms is considering paid agreements with news publishers to access high-quality content and train its AI models.
- This initiative aims to enhance the efficacy and competitiveness of Meta’s generative AI tools, such as Meta AI, by utilizing premium news articles, photos, and videos.
Facebook’s parent company, Meta, is said to be in talks with news organizations to improve the training of its artificial intelligence (AI) language models. Meta is looking to build up its generative AI tools, like Meta AI, to rival Google and Microsoft offerings. According to insiders, internal discussions about accessing news, photo, and video content are being held, but no formal approach has been taken yet.
Mark Zuckerberg had previously claimed that Llama AI models of Meta use more data than the Common Crawl dataset. However, concerns over quality persist within its ranks. For training generative AI, it is important that high-quality content such as books or news articles exists to drive better outputs, as social media posts and comments may not have the same level of training value.
OpenAI also recently signed a deal with News Corp worth $250m over five years, allowing them to access content from well-known outlets such as the Wall Street Journal and MarketWatch. This is an example of an increasing trend where artificial intelligence firms seek top-notch datasets to refine their models. Thus, Meta may be on the verge of working directly with publishers soon, having mused internally about it.
See more: Slack Caught Using User Data To Train AI Models Without Explicit Permission
Conclusion
These possible partnerships between Meta and news publishers indicate a shift towards increasing the quality of AI training data, similar to OpenAI’s strategy. In this race for artificial intelligence, obtaining high-quality materials could be instrumental in designing more effective generative AI tools.
MORE ON ARTIFICIAL INTELLIGENCE
