Google has struck a deal worth $60 million that will allow it to use Reddit content to train its generative-AI models, Reuters reported on Thursday, citing three people familiar with the matter.
The claim follows a Bloomberg report earlier in the week that said Reddit had inked such a deal, though at the time, the name of the other party remained unclear.
Training AI models using human-generated content such as that found on Reddit is supposed to help chatbot tools respond in a more natural and conversational way, and with relevant and up-to-date information.
Reuters’ report comes as AI companies search for new ways to utilize large amounts of online data without upsetting those who own the copyright to it. It also comes as Reddit announced plans for its initial public offering, in which it said it will list its shares on the New York Stock Exchange under the ticker symbol RDDT.
Up to now, most AI models that underpin tools like OpenAI’s ChatGPT or Google’s Gemini (formerly Bard) have been trained mainly on content scraped from the web. But the method has caused alarm among authors, artists, publishers, and others, as their copyrighted work is used without any form of recognition or, more importantly, financial compensation. Some are taking the AI companies to court over copyright infringement, a situation that’s prompted the AI tech firms to explore new ways to obtain content, such as deals with sites like Reddit that hold vast amounts of useful material.
Reddit’s reported deal with Google echoes one made recently by Axel Springer that gives OpenAI access to the German media giant’s content for AI training — though this approach also has its issues. For example, some have expressed concern that such deals will see the money go straight into company coffers rather than shared among those creating the content.
A Wired article in December highlighted the matter in relation to the Axel Springer deal, saying that it was “entirely unclear whether individual journalists will see any of that money. When asked if reporters would benefit from any revenue-sharing or additional compensation as a result of the licensing arrangement, Axel Springer did not directly answer the question … So, as of now, it is unclear whether a writer whose work is incorporated into ChatGPT will receive a one-time payment, a recurring royalty-like payment, or no payment at all.”
Reddit and Google both posted announcements on Thursday describing a move toward closer collaboration in a number of areas, though neither directly mentioned the recently reported deal or its value.
Google said Reddit contained “an incredible breadth of authentic, human conversations and experiences” while Reddit commented that its work with Google “will make it easier for people to find, discover, and engage in content and communities on Reddit that are most relevant to them.”
