OpenAI, the maker of the viral AI chatbot ChatGPT, has signed a new news licensing deal in Europe, adding London's Financial Times to a growing list of publishers paying content access fees.
As with previous OpenAI publisher license agreements, the financial terms of the deal were not made public.
The latest deal looks a bit more cosy than other recent OpenAI publisher partnerships, including those with German giant Axel Springer and France and Spain's Associated Press, Le Monde Press, and Prissamedia. The two companies called this arrangement a “strategic partnership and license agreement.” (However, Le Monde's CEO also referred to the “partnership” with OpenAI announced in March as a “strategic move.”)
However, we understand that this is a non-exclusive license agreement and OpenAI does not acquire any type of equity interest in FT Group.
In terms of content licensing, the pair said the agreement covers the use of FT's content in training AI models with OpenAI and, if necessary, displaying it in generated AI responses generated by tools such as ChatGPT. “This is similar to other publisher agreements,'' it said.
The strategic elements appear to be particularly focused on advancing the FT's understanding of generative AI as a content discovery tool, with the aim of developing “new AI products and features for FT readers”. What is being touted as a collaboration suggests news publishers are keen. To expand the use of AI technology more generally.
“Through the partnership, ChatGPT users will be able to see selected summaries, quotes and rich links to FT journalism in response to relevant questions,” the FT said in a press release.
The publisher also noted that it became a customer of OpenAI's ChatGPT Enterprise product earlier this year. They also signaled that they want to explore ways to further deepen their use of AI, while expressing caution about the potential risks to the reliability of automated output and reader trust.
“This is an important agreement on many counts,” FT Group CEO John Ridding said in a statement. “This will recognize the value of our award-winning journalism and provide early insight into how content surfaces through AI.”
He continued: “Apart from the benefits to the FT, there are broader implications for the industry. Of course, it is right for AI platforms to pay publishers for the use of their material. , we understand the importance of remuneration. These are all essential to us. At the same time, it is clearly in the interest of our users that these products include reliable sources of information.”
ChatGPT Large-scale language models (LLMs) such as OpenAI's GPT that power chatbots are notorious for their ability to fabricate information or “hallucinate.” This is the opposite of journalism, which verifies that the information reporters provide is as accurate as possible.
So it's actually not surprising that OpenAI's early moves toward licensing content for model training were centered around journalism. The AI giant may hope this will help solve the “hallucination” problem. (A line in the PR suggests this partnership will “help improve”) [OpenAI’s] Learn the usefulness of the model by learning from FT journalism. ”)
However, there is another big motivating factor here as well. That is a legal responsibility regarding copyright.
Last December, the New York Times announced that it was suing OpenAI, alleging that copyrighted content was used by the AI giant to train models without a license. OpenAI believes that the only way to limit the risk of further litigation by news publishers is that content collected (or has been collected) from the public internet to aid in the development of LLM is protected by copyright. It is disputed that the publisher is required to pay a fee for the use of the published content.
Publishers stand to earn large sums of cash from content licenses.
OpenAI told TechCrunch that “approximately a dozen” publisher deals have been signed (or “imminent”), adding that “many” more deals are in the works.
Publishers can also potentially gain additional readers, such as when users of ChatGPT choose to click on quotes that link to their content. However, over time, generative AI could also cannibalize search engine usage and divert traffic away from news publishers' sites. With this kind of disruption on the horizon, some news publishers may see a strategic advantage in forging closer relationships with the likes of OpenAI.
There are some reputational pitfalls when engaging with Big AI, even for publishers.
Technology publisher CNET rushed to introduce generative AI as a content creation tool last year, but without fully revealing to its readers how it would use the technology, a journalist at Futurism wrote an article about machine-written articles. Its reputation suffered even more when numerous errors were discovered in the book. was published.
The FT has a reputation for producing quality journalism. So it will certainly be interesting to see how the company further integrates generative AI into its products and news editing processes.
Last month, the company announced its GenAI tool for subscribers. This is essentially a transition to providing natural language search options for 20 years of FT content (so it's essentially a value-add aimed at driving subscriptions to human-produced journalism) .
Furthermore, due to legal uncertainty in Europe, many privacy law concerns prevent the use of tools such as ChatGPT.