OpenAI and Google have announced new collaborations with major publishers to expand access to information for their AI products. OpenAI is working with Time magazine over several years to give its machine learning technology access to the publisher’s assets for improved AI tools and support for authoritative journalism. The integration of Time’s content into OpenAI’s products may involve the use of Recall Augmented Generation to generate more accurate answers to user queries.
Google has also expanded its content partnerships with companies like Moody’s, MSCI, Thomson Reuters, and Zoominfo. These collaborations will provide data for Google’s RAG databases to improve the quality of AI-generated results. However, concerns have been raised about the quality of source data and the potential for errors if inaccurate information is used.
Both OpenAI and Google have faced criticism over AI systems providing broken URLs for publications that have licensing deals with them. OpenAI’s partnerships with publishers like Time and Reddit have aimed to provide proper attribution to original sources but have sometimes failed to deliver on this promise. Additionally, the Center for Investigative Reporting has accused OpenAI and Microsoft of using copyrighted content without permission or compensation.
Overall, these partnerships highlight the growing importance of high-quality data sets for machine learning tools and the need to protect the integrity of content sources in the AI landscape. The collaborations between tech giants and publishers aim to enhance AI capabilities while respecting copyright and attribution principles. It will be essential for these companies to address concerns around source quality and provide transparency about their data partnerships moving forward.
Article Source
https://www.theregister.com/AMP/2024/06/28/openai_google_ai/