Several factual Mistakes Were Made in The Launch Demo of Microsoft’s Bing A.I.
Last week, chatbots were all the rage. Microsoft And Google In an effort to compete with each other, over 1 million people signed up for Microsoft’s early versions of AI-powered search within the first 48 hours.
Satya Nadella, Microsoft CEO, about the technology that can give complete answers that look like they were written by humans.
However, accuracy is a concern for many.
Microsoft’s Demo In front of reporters, ChatGPT-like technology embedded within the company’s Bing search engines analyzed earnings reports Gap And Lululemon . The chatbot did not match the reports with its answers. Some numbers may have been fabricated.
“Bing AI got some answers totally wrong during their demo. Dmitri Brereton, an independent search researcher, wrote in a Substack article on Monday. Instead, everyone got on board the Bing hype train.
Breton pointed out possible facts in the Microsoft demo’s responses to questions about vacuum cleaner specifications, travel plans to Mexico, and financial errors. CNBC reported that he didn’t look for errors at first, but only later did he notice them in his attempt to compare the AI announcements from Google and Microsoft.
AI experts refer to the phenomenon as “hallucination,” which is the tendency of large-language models to make things up. Google released a rival AI tool last week that included factual errors, although they were quickly pointed out by viewers.
Both companies are eager to integrate new types of generative AI into search engines. This is in response to ChatGPT’s November introduction to the public. OpenAI has raised billions of dollars from Microsoft while rival startups such as Hugging Face and Stability AI have also risen to billion dollar valuations through private funding rounds.
