Apple Looks for News Media Partners to Train Generative AI Models: Report
Apple wants to train future generative AI models with news media archives and is negotiating a deal with major publishers to secure access, according to report by The New York Times. These negotiations involve licensing deals potentially worth over $50 million for each publisher, signifying Apple’s growing focus on large language models (LLMs) and generative AI after largely being a secondary player in the space this year among tech giants.
Apple Generative AI
The planned acquisition of decades of news content would provide Apple’s AI systems with a vast repository of data, crucial for training sophisticated machine learning models. Publishers approached by Apple for multi-year agreements include the likes of Condé Nast, NBC News, and IAC. The talks are part of a larger initiative from Apple to enhance its generative AI research and development. Presumably, the long-term goal is to make Apple a competitor with the brands increasingly identified with generative AI, both the newer names like OpenAI and Anthropic, as well as fellow industry behemoths investing heavily in LLMs, such as Microsoft and Google. Apple’s recent AI efforts have centered around projects like optimizing models for Apple Silicon, Siri upgrades, and the integration of generative AI-powered features in upcoming versions of iOS.
For Apple, gaining access to news archives could significantly accelerate its AI development. Meanwhile, for publishers, the decision to enter these deals involves carefully balancing the potential risks and rewards with far-reaching implications for their digital future. The plans aren’t finalized, however, and the report cited some caution from publishers about the project. Concerns revolve around potential legal liabilities and the unforeseen capabilities of generative AI models developed using their content. Additionally, there is apprehension that Apple might eventually leverage this access to compete directly with the news industry, with the company being ambiguous about its long-term intentions. Still, the publishers are likely to prefer that to the content scraping that has prompted many lawsuits against LLM developers. The litigious backlash prompted OpenAI to strike agreements like the ones Apple wants with publishers such as the Associated Press. OpenAI also just announced a deal with Berlin-based publishing giant Axel Springer to integrate articles from its publications into ChatGPT, providing answers to users and a data pipeline for training OpenAI’s LLMs.