Google warns about ‘hallucinating’ chatbots

1
Google warns about hallucinating chatbots
In a newspaper interview published on Saturday,
CEO of Googles search engine warned against the
pitfalls of artificial intelligence in chatbots,
as Google parent firm Alphabet struggles to
compete with blockbuster software ChatGPT. This
kind of artificial intelligence we are talking
about right now might sometimes lead to
something we call hallucination, Prabhakar
Raghavan, Googles senior vice president and head
of Google Search, told the German daily Welt am
Sonntag. This then expresses itself in such a
way that a machine produces a persuasive but
entirely fabricated answer, Raghavan wrote in
remarks published in German. One of the most
important objectives, he continued, was to keep
this to a minimum. Google has been on the
defensive since OpenAI, a startup backed by
Microsoft for approximately 10 billion, debuted
ChatGPT in November, which has subsequently
astonished users with its startlingly human-like
responses to user queries. Alphabet Inc launched
Bard, its own chatbot, earlier this week, but the
software incorrectly revealed information in a
promotional video, costing the business 100
billion in market value on Wednesday. Alphabet,
which is still performing user testing on Bard,
has not announced a release date.
2
We certainly feel the urgency, but we also feel
a tremendous amount of duty, Raghavan explained.
We do not want to mislead the people.