OpenAI Strikes ChatGPT Content Deal With Conde Nast

In a blog post late yesterday, OpenAI announced a brand new partnership with global publisher Condé Nast. It sounds like a simple win-win for both companies: OpenAIl gains access to a huge archive of back-issue content from well known titles like Vogue, GQ, and The New Yorker, among its 20 or so print and digital titles, which it can use to train its AI systems. Condé Nast gets its publications’ stories “surfaced” in ChatGPT’s systems when users ask relevant questions of the AI.

Who wins?

Though the exact details of the deal were kept under wraps, Condé Nast’s CEO sent a memo to staff that implied that the deal would last several years and involve some form of payment from OpenAI, according to news site TechCrunch. “It’s crucial that we meet audiences where they are and embrace new technologies while also ensuring proper attribution and compensation for use of our intellectual property,” Roger Lynch said in the memo, adding “This is exactly what we have found with OpenAI.” 

For its part, OpenAI revealed that it would be displaying Condé Nast content in both ChatGPT and its experimental SearchGPT system. In terms of the search feature, OpenAI’s blog post explains it’s been “testing new search features that make finding information and reliable content sources faster and more intuitive.” To do this it’s been combining its chatbot-style AI system with “information from the web to give you fast and timely answers with clear and relevant sources.” It also notes in the future these will be integrated into ChatGPT, when the “SearchGPT” experiment ends. Most interestingly, OpenAI also points out it’s been working directly with its news partners to improve how SearchGPT works from their point of view.

The future of AI search?

These last couple of details may be the most interesting bit of this news. We know OpenAI has been slowly signing up content producer partners such as social forum chat site Reddit, and the huge media empire of News Corp. But the fact that in its new deal with Condé Nast, OpenAI has demonstrated it’s already improving its search skills, and is accepting input on the process from its publishing partners will be interesting to Google critics and ad industry experts. It’s more proof that OpenAI is directly aiming at stealing search traffic from Google–which has been experimenting, controversially, with AI search itself. When OpenAI embraces search directly within ChatGPT, it’s possible that the entire online ad industry will take a hit, since Google dominates the market. The resulting wobbles could hit the many thousands of small companies that rely on Google’s ad system for income in one way or another.

The fact that an outfit like Condé Nast could be helping OpenAI hone how AI search works could also be read as proof this isn’t a “deal with the devil” for the publishing giant, which will have one eye on protecting access to its revenue-generating content archive and another on protecting future income streams in an era when generative AI can spew out text and image content that easily mimic the genuine human-made material from content publishers.

Deals over scraping data illegally

This new content licensing deal occurs against a background of splashy scandals that have seen several “Big AI” brands, including OpenAI itself, accused of unfairly acquiring megatons of other people’s intellectual property to train their AI algorithms. Just this week the San Francisco-based AI startup Anthropic was sued by three writers who accused the AI maker of stealing lots of other writers’ material at a “large scale” to train its Claude AI. The Condé Nast deal stands in contrast to this. Will it blaze a trail for AI outfits like OpenAI to sign more and more deals with content providers, big and small? Will individual influencers broker deals with OpenAI to sell access to their content, and ensure that their new material is “surfaced” in ChatGPT? Time will tell.