Google should be used as platform to cross-check information from Bard AI: Google UK Executive

economictimes.indiatimes.com:

Google countered OpenAI's chatbot ChatGPT with the release of Bard, and the recent report by Netskope also suggests that Google’s AI chatbot has been attracting new users with ease lately. The report highlights that Google's Bard has been adding users at a rate of 7.1 per cent a week, whereas rival ChatGPT struggles at 1.6 per cent. This makes Google Bard the first fastest-growing AI app in current times.

In the case of ChatGPT’s free version, OpenAI has been quite clear about the limitations and its «knowledge cut-off» which is till 2021. Now, Google’s UK Managing Director, during an interview with BBC, made a revelation that users should always cross-check Bard’s answers. Google's UK boss Debbie Weinstein told BBC, Bard was not a type of place where users can go and look for «specific information».

She said that users can use Google as the search engine in order to «actually reference information they found.» This doesn’t come unexpectedly, as Bard's homepage reads that it may «display inaccurate or offensive information», and now Google UK executive has also warned users to not solely rely on the information that Bard produces. «Bard is an experiment and may give inaccurate or inappropriate responses. You can help make Bard better by leaving feedback,» a message on Bard's homepage reads.

The reasoning behind this is hallucination, which refers to the model confidently providing inaccurate information. Even OpenAI's powerful GPT-4 large language model is not exempt from this issue. During its launch in February, Bard demonstrated incorrect responses, causing a significant drop in the company's share price.