Amidst the wave of ChatGPT bots, Vint Cerf, dubbed the “Father of the Internet,” has warned business leaders against rushing to invest in conversational AI technologies just because they are “really cool.”
"We have an ethical issue here that I hope some of you will consider," Cerf told the audience at the Mountain View conference on Monday. Everyone talks about (GBT chat) or a copy of Google and we know that it doesn't always work as expected.
Cerf's warning comes as big tech companies like Google, Meta and Microsoft struggle to stay competitive in conversational AI, a technology whose rapid rise continues to make huge mistakes.
Far from widespread, AI chatbots have been linked to bugs and related issues, to an unknown degree of "toxicity," said John Hennessy, president of Alphabet (Google's parent company), earlier Monday. It must be resolved before being released to the public.
Cerf has been Vice President of Google and Chief Internet Advocate since 2005. He is known as one of the "Fathers of the Internet" because he was involved in designing some of the structures that laid the foundations for the Internet.
Cerf cautions against investing in technology just because it's "really cool," even if it doesn't always work out. He also warned companies against greed, saying, "They will try to do what is good for them, not for you." "So we have to keep that in mind and think seriously about how we use this technology."
Cerf said he tried to get an AI to append symbols at the end of each sentence but failed, and when he remembered he apologized but didn't change his behaviour.
(Serve) gave another example, where he asked the robot for a biography about himself, and the robot returned the resume as if it was a proven fact, even though the resume contained inaccurate information.
"From an engineering standpoint, I think it's up to engineers like myself to find a way to tame some of these technologies so that they're less likely to cause harm," Cerf said.
He added, "Of course it depends on the application, a not-so-good tale is part of what is expected of a robot. But giving advice to someone could have medical consequences. Knowing how to reduce the possibility of the worst is very high."