![](https://www.techtarget.com/rms/onlineimages/security_a244600171_searchsitetablet_520X173.jpg)
Alex - stock.adobe.com
OpenAI CondeNast SearchGPT deal addresses copyright concerns
The AI vendor continues to expand its SearchGPT prototype by partnering with news publisher Conde Nast. It also introduced fine-tuning for its latest model.
OpenAI has added another major publisher to its list of partnership deals and updated its enterprise-ready offerings and applications this week.
On Tuesday, the ChatGPT maker revealed that it partnered with news publisher Condé Nast to display content from brands such as The New Yorker, Vogue, GQ, Architectural Digest and Vanity Fair within ChatGPT and its search prototype SearchGPT.
The AI vendor also launched fine-tuning for GPT-4o, claiming that the feature is the most requested from developers. It will also offer one million training tokens for free for every organization until September 23.
The partnership and copyright infringement
The multi-year Condé Nast partnership fulfills a promise from OpenAI to partner with more news publishers as it prototypes SearchGPT.
OpenAI has also partnered with The Atlantic, Axel Springer, and Vox Media, as well as NewsCorp, which owns The Wall Street Journal and The New York Post.
The deal also coincides with an ongoing lawsuit between the AI vendor and news publisher New York Times. The vendor is not the only one accused of infringement. Its competitor, Anthropic is was also sued by authors for allegedly infringing on copyrighted books.
The deal between OpenAI and Condé Nast is smart for both sides, according to industry experts. The deal ensures Condé Nast publications receive proper attribution and compensation when used by OpenAI. For OpenAI, a deal like this helps in dealing with copyright infringement concerns, said Michael Bennett, AI policy Adviser at Northeastern University.
"This type of licensing deal probably makes more sense to industry players as an alternative to bringing an expensive lawsuit," he said.
These types of deals will likely become routine and help combat the problem of AI-generated content that contains misinformation, Futurum Group CEO and analyst Lisa Martin said.
"I believe it is a positive step to truthful content getting to end users faster, and training more accurate LLMs for the greater good," Martin said. "We have to do things like this to positively benefit society."
A deal like this also helps OpenAI’s SearchGPT prototype, Bennett said.
For one, a ChatGPT augmented search gives OpenAI an edge over Google, he said. Also, OpenAI’s competitors looking to copy the SearchGPT model will be "boxed out by exclusive or de facto exclusive licensing deals" between OpenAI and news publishers, he said. However, Conde Nast’s deal with OpenAI is not exclusive.
It’s also too soon to tell if the partnership approach is sustainable for publishers, Martin said.
"We don't know yet if what companies like OpenAI are paying is sufficient for the publishers," she said. She added that as deals like these become common, publishers will likely ask for more money. "We also don't know how this model will affect the future of ad-supported journalism or if this model will enable publishers to thrive," she continued.
Moreover, it’s too soon to determine how OpenAI will make use of this partnership, Gartner analyst Arun Chandrasekaran said.
"Whether this partnership will extend beyond search into broader model training for OpenAI is yet to be seen," Chandrasekaran said.
Fine-tuning for 4o
While partnering with publishers will help OpenAI succeed in steering users to its searchGPT and ChatGPT products, the AI vendor also aims to steer developers to its GPT-4o model.
While responding to developers’ demands, OpenAI is also evaluating its competitors and what it takes to be relevant in the market. OpenAI is also thinking about competitors in its decision to launch a new fine-tuning feature for GPT-4o.
"They’ve got to more or less keep up with what the market wants," said Mark Beccue, an analyst with TechTarget’s Enterprise Strategy Group.
Fine-tuning allows enterprises and organizations to use their data, and many want to do that, he said.
Some of OpenAI’s competitors, including Anthropic, offer fine-tuning. Thus, it’s only a matter of time for OpenAI to do the same for its 4o model.
"These are the natural progressions of all these models and what they are capable of," Beccue continued. "It’s not the end all be all. It’s one of several customization tools."
While fine-tuning is important, most enterprise developers will continue to rely on prompt engineering and retrieval augmented generation, Chandrasekaran said.
Esther Ajao is a TechTarget Editorial news writer and podcast host covering artificial intelligence software and systems.