Google says Bard users should fact-check responses in search engine
Google says its AI chatbot, Bard, is prone to making mistakes and urges users to fact-check responses via its search engine.
Google’s UK Vice President Debbie Weinstein told the BBC’s Today Programme on Friday that Bard should primarily be viewed as a creative tool capable of “problem solving”.
She noted that Google Search is still better for finding specific information and factual content.
By referencing responses in its search engine, Weinstein also advised users to double-check whether Bard has made a mistake.
The surprising admission by Google highlights the limitations of its AI-powered conversational chatbot.
Bard launched several months ago in the wake of ChatGPT’s incredible success.
It remains in an “experimental” phase and the main page states that Bard “won’t always get it right”.
However, Weinstein’s comments suggest it might be incapable of basic search tasks.
Google’s boss advised users to click the feedback buttons to help train Bard.
But users will be hopeful that further updates can improve the quality and integrity of responses.
That’s an important objective as interest in AI surges.
A Google report found search queries involving AI hit a record high in the first six months of 2023.
The most popular search term was “how to make money with AI”.
Google fully understands the importance of that query after it sounded an internal “code red” amid concerns about how AI could affect its lucrative search engine business.
UK Prime Minister Rishi Sunak is also excited about the economic opportunities artificial intelligence could unlock in the coming years.
However, many experts have raised concerns about how AI could potentially reshape labour markets and humanity as we know it.