Like Google’s artificial intelligence (AI) chatbot Bard, the new AI-powered Bing makes obvious mistakes in its first trials. These artificial intelligence-based search engines, which are full of mistakes for now, also defend their mistakes as if they were true.

Independent AI researcher Dmitri Brereton discovered that the financial data Microsoft showed in demos of its new Bing search engine was full of errors. Bing’s artificial intelligence appears to have mixed data from Gap Inc and Lululemon. Bing also makes an obvious mistake in a query about the pros and cons of its best-selling vacuum cleaners for pet hair. In this error, he mentions that the product has a short cable, whereas the product is wireless. There is also a variant with a cable of approximately 5 meters. It seems that Bing’s AI is confusing the two products.

Bing: “Today is 2022”

Bing’s bugs aren’t limited to the demos, of course. As it is known, Microsoft has created a waiting list for the new Bing and is testing the process on a limited audience. According to a reddit user, Bing is having trouble with today’s date.

The user asks Bing for today’s date and gets the answer February 12, 2023. In the ongoing conversation, the user says that Avatar 2 will be released then, but Bing is confused and refers to 2022 as the year that comes after 2023. Afterwards, Bing admits that he made a mistake and says, “If we are in 2023, then 2022 is not in the future” and adds, “But we are not in 2023, we are in 2022”. Although Reddit user tries to fix this error, Bing says, “Please trust me, I’m Bing and I know the date.”

Although the user has complained about the problem, Bing says that the problem is caused by the user’s phone, the date of the phone may be incorrect, and it may be caused by a virus.

Microsoft is aware of the problems

Microsoft, on the other hand, is aware of the problem. Caitlin Roulston, director of communications, emphasizes the importance of feedback, saying, “We expect the system to make mistakes during this preview period, and feedback is critical in helping us identify where things aren’t going well so we can learn and help models get better.” The new Bing isn’t just about bad answers, of course. Besides the creative answers, the ability to continue the conversation is also better than ChatGPT. In fact, unlike ChatGPT, Bing ends the conversation when you ask it the same question over and over. If you’re still on the waiting list for the new Bing, you can see what others are up to on the Reddit page here.

However, the bugs of the new Bing are not limited to these. Microsoft already says at the heart of its search engine that it uses a more specific and complete version of the model used in ChatGPT. But ChatGPT makes tons of mistakes and can make them look real. As a result, AI-powered search engines seem to have a long way to go in terms of trust and accuracy.