ChatGPT introduces its own AI search engine that combines the power of AI models with up-to-date information from the web. SearchGPT is in the prototype phase, but users can join a waitlist to test it out.

OpenAI’s new AI search engine “SearchGPT” combines the power of OpenAI’s AI models with current information from the web. The company apparently wants to respond to the various allegations related to the use of copyrighted information. Anyone who would like to try out SearchGPT can register for a waiting list on the ChatGPT website.

New way of searching

Just as Google searches the internet to find an answer to your search query, so do AI models. ChatGPT searches the internet for text to formulate an answer to your question. “What if you could combine the two?” That’s what they thought at ChatGPT.

The new AI search engine SearchGPT is currently in the testing phase. It is a prototype of new search capabilities that combines the power of OpenAI models with information from the web. This should increase the speed of ChatGPT’s responses combined with relevant and clear sources.

AI models are trained on data, but this data can become outdated after a certain period of time. If you ask who the current Belgian participants in the Olympics are, ChatGPT cannot provide an up-to-date answer, unlike the search engine. The prototype tool is “temporary,” but interested users can join the waiting list to try out SearchGPT.

Kingship

OpenAI has been criticized several times for training its models on copyrighted information without the authors’ permission. With SearchGPT, the company seems to be making its awareness of this clear.

“We’ve worked with publishers to build this experience and continue to gather their feedback,” OpenAI said. The search engine could be an important step in copyright discussions because it is “a way for publishers to manage how they appear in it.”