OpenAI has announced SearchGPT, an AI-powered search engine. The startup said that the new platform would be able to respond to search queries conversationally with up-to-date information from throughout the internet while citing its sources. OpenAI said that SearchGPT is launching to a small group of users and publishers ahead of its planned integration into ChatGPT.
“Getting answers on the web can take a lot of effort, often requiring multiple attempts to get relevant results,” said OpenAI. “We believe that by enhancing the conversational capabilities of our models with real-time information from the web, finding what you’re looking for can be faster and easier.”
SearchGPT a clear challenge to Google
SearchGPT marks OpenAI’s first foray into AI-powered search engines – a market that, until now, has been slow to gain traction among mainstream users. Though several startups debuted similar search engines last year, user uptake has remained low. Underlying models, meanwhile, continue to be plagued by error-strewn results and hallucinations, including add-ons to conventional search engines debuted by more established market players like Microsoft Bing. Perhaps most importantly, any new entrants into the search market find themselves competing against Google, which effectively dominates the sector.
This does not seem to have fazed OpenAI, which pitched SearchGPT’s in-built connection to an ecosystem of pre-approved publishers as a unique selling point for the new platform. Results from these outlets are prioritised in the new search engine, explained its creator, with responses having “clear, in-line, named attribution and links so users know where information is coming from and can quickly engage with even more results in a sidebar with source links.”
The startup revealed that one of these partners was the US magazine The Atlantic – a fact lauded by its chief executive, Nicholas Thompson. “AI search is going to become one of the key ways that people navigate the internet,” said Thompson. “It’s crucial, in these early days, that the technology is built in a way that values, respects, and protects journalism and publishers.”
OpenAI balancing act with publishers
OpenAI added that it is working on new ways for publishers to control how their content appears on SearchGPT as well as improving user experience “in areas like local information and commerce.”
OpenAI’s relationship with media publishers overall has been somewhat mixed recently. In December 2023 the New York Times filed a civil suit against the startup, claiming that it had trained ChatGPT and other models on its articles without permission. This was followed by another lawsuit mounted by eight US newspapers, including the Chicago Tribune and the Denver Post, in April of this year. Other publishers have been more open to allowing OpenAI to use their content as training data. In May, the startup announced a series of partnerships with News Corp, The Atlantic and Vox Media permitting its developers to access their archives.
“We believe [this] historic agreement will set new standards for veracity, for virtue and for value in the digital age,” said News Corp’s chief executive, Robert Thomson. “This landmark accord is not an end, but the beginning of a beautiful friendship in which we are jointly committed to creating and delivering insight and integrity instantaneously.”
Read more: Microsoft abandons observer seat on OpenAI board
