In May, rumors started hitting the internet that OpenAI was about to launch its own search engine as part of its spring media update on May 13. Those reports were quickly shot down by OpenAI CEO Sam Altman. Instead, the spring event was centered on the reveal of the GPT-4o AI model.
Today, OpenAI officially revealed it is working on its own internet search engine, but in a far more low-key way. The company"s official website has the details on the search engine, which is naturally called SearchGPT.
The blog post states:
SearchGPT will quickly and directly respond to your questions with up-to-date information from the web while giving you clear links to relevant sources. You’ll be able to ask follow-up questions, like you would in a conversation with a person, with the shared context building with each query.
The blog post adds that SearchGPT is designed to help internet-based publishers get more hits for the content that will come up as answers to questions with "clear, in-line, named attribution and links" to those sources. The post added that sites that block their content from being used to train generative AI models like ChatGPT can still allow their information to be used as sources for SearchGPT.
People who want to try out SearchGPT can do so by heading to its official page and sign up for the waitlist. There"s no word on how many people will be picked to check out the search engine. OpenAI says that eventually, features that are tested in the SearchGPT prototype will be added to its ChatGPT AI model.
SearchGPT was revealed one day after Microsoft announced it was adding a new AI generative model for a limited number of its Bing search engine users. It also comes a few months after internet search leader Google announced it was rolling out AI Overview answers for questions in its search engine.