OpenAI's SearchGPT appears to get lost on its first hunt
SearchGPT hallucinates answer in first demo video
![OpenAI](https://cdn.mos.cms.futurecdn.net/BHjXPtvs6sK9QX4XiCPrHM-320-80.jpg)
The new SearchGPT feature introduced by OpenAI has stumbled in its rollout, as first discovered by The Atlantic. SearchGPT is a new tool designed to combine OpenAI's AI models with real-time web data for faster, more accurate answers. It's not widely available but represents OpenAI's vision of how AI and search will augment each other in the future.
In a prerecorded video showcasing SearchGPT, the tool provided incorrect information about the dates of the Appalachian Summer Festival in Boone, North Carolina. The July 29 to August 16 dates SearchGPT reported are far from the June 29 to July 27 dates the event actually occurs. As reported by The Atlantic and further confirmed, the dates of the festival in SearchGPT's response are when the box office is closed. You can watch the full demo here.
SearchGPT Still Looking
AI hallucinations and errors are perennial complaints, a universal issue pretty much every AI user has encountered. In that sense, the mistake is not a huge deal. However, the result here somewhat undermines OpenAI's pitch for SearchGPT. With reliability and transparency at the heart of SearchGPT, according to OpenAI, it stands out when a hallucination like this still happens.
This isn't an issue that is unique to OpenAI, of course. You may recall the embarrassing errors made when Google debuted its AI assistant Bard (now Gemini). In that case, a live demo claimed the James Webb Space Telescope took the first pictures of a planet outside the solar system, even though they were actually taken by the European Very Large Telescope. It was so bad that some people attributed a subsequent $100 billion dip in Google stock price to the gaffe. Arguably, having the error in a video that OpenAI could have redone or edited differently at any time is worse than a real-time misstep.
Regardless of the mistake, OpenAI will likely not slow down in developing AI search. The demand for accurate, speedy answers to questions about things happening right now has already propelled AI integration into existing search engines, including Google and Microsoft Bing. It's also the impetus behind search-centered generative AI chatbots like You.com and Perplexity.
But, if OpenAI can release its own option, especially tied to ChatGPT, it will, at the very least, score a robust slice of the search AI market. That's predicated on getting people to trust SearchGPT to provide accurate answers, of course. You could ask SearchGPT to give the odds of that happening, but you might want to follow up on your own to confirm what it says.
You might also like...
- OpenAI unveils SearchGPT and prepares to take on Google
- Perplexity's AI search could eliminate the need for follow-ups and beat ChatGPT at its own game
- OpenAI may be working on AI that can perform research without human help – which should go fine
Get daily insight, inspiration and deals in your inbox
Sign up for breaking news, reviews, opinion, top tech deals, and more.
Eric Hal Schwartz is a freelance writer for TechRadar with more than 15 years of experience covering the intersection of the world and technology. For the last five years, he served as head writer for Voicebot.ai and was on the leading edge of reporting on generative AI and large language models. He's since become an expert on the products of generative AI models, such as OpenAI’s ChatGPT, Anthropic’s Claude, Google Gemini, and every other synthetic media tool. His experience runs the gamut of media, including print, digital, broadcast, and live events. Now, he's continuing to tell the stories people want and need to hear about the rapidly evolving AI space and its impact on their lives. Eric is based in New York City.