OpenAI Nets FT London, Wake up Google! 

More than a month ago, OpenAI had bared its fangs and let competitors know that they were ready to pay good dollars to media houses that would give them access to data needed to build AI models. While the battle with New York Times is still on, the Microsoft-funded company has now netted Financial Times of London in another non-exclusive deal. 

And true to form, OpenAI has kept the licensing deals under wraps as it had done earlier when they struck similar ones with the likes of Le Monde and Prisa Media to bring French and Spanish news to their platform. However, market watchers believe that the latest deal could be the best of such publisher tie-ups. 

The company, which already has deals with German media company Axel Springer and Associated Press, is now calling its alliance with the Financial Times as a “strategic partnership and licensing agreement” unlike in the past where it was usually mentioned as a strategic move by the company, which was funded $10 billion by Microsoft. 

The deal with FT seems better phrased

However, OpenAI seems to have taken care to not make it an exclusive arrangement and fall foul of antitrust regulators in Europe. Thus, the duo have agreed to let the ChatGPT founders to use FT’s content for training AI models while also displaying GenAI responses as appropriate when their chatbot is queried by users. 

Now, isn’t that what OpenAI did with its other deals too? In which case, what’s this strategic alliance that they’re talking about? This is where Google should get a bit concerned as the deal also involves FT boosting its knowledge of GenAI as a content discovery tool that would be rolled out for its readers. 

“Through the partnership, ChatGPT users will be able to see selected attributed summaries, quotes and rich links to FT journalism in response to relevant queries,” says FT in a press release. It also noted that FT acquired ChatGPT Enterprise licenses for all its staffers in order to ensure than they were “well-versed in the technology and can benefit from the creativity and productivity gains made possible by OpenAI’s tools.” 

Media houses need money, and a Google breaker

Group CEO John Ridding launched into hyperbole claiming that the alliance was a recognition  of the company’s “award-winning journalism and will give us early insights into how content is surfaced through AI.” Quite clearly, there is more to this than data modelling and that something could be quite a concern for Google. Are you listening? Sundar Pichai. 

Of course, Ridding immediately sought to allay fears of AI removing humans from journalism. He waxed eloquent about FT’s commitment to human journalism while also giving a thumbs up for OpenAI’s understanding of “transparency, attribution, and compensation” which are all essential for FT as a media group. 

OpenAI is playing its cards well

As readers would be aware, large language models (LLMs) are the engines that power chatbots such as ChatGPT which makes up information or hallucinates as was seen in the past. How such an engine can join hands with a media brand where reporters need to verify each and every bit of information for accuracy is anybody’s guess. 

From OpenAI’s point of view (and for competitors as well), licensing content to train AI models around journalism is the best move they can make. This would help them fix, if not completely, the challenges around chatbots making up information. Small wonder then that the press release refers to the alliance as FT’s attempt “to help improve models’ usefulness by learning from FT journalism.” 

We aren’t even referring here to the issue of legal liability as OpenAI is already battling a lawsuit with New York Times, which has accused it of using copyrighted content to train models. And doing so without a license. While OpenAI is battling this case, it has sought to pay media houses for content to train its models now. 

Will Google be left high and dry on search?

As for publishers, they stand to gain some hard cash for such licensing deals as well as the promise of new subscribers should users of ChatGPT click on citations that link to their content. And this is what Google should be concerned about as it could drive away search traffic from news websites. 

On its part, Google has made a slow start by suggesting that SEO efforts led by chatbot-led content writing could face demerit points in their search rankings. However, if OpenAI is ready to use big chunks of its cash reserves to pay off reputed publications for licensing their content, will such a move alone suffice? We think not!