The massive popularity of the generative A.I. bot ChatGPT has led tech companies to release their own versions of the tool, from Google’s Bart to Chinese giant Baidu’s Ernie Bot. Some schools are even moving to include ChatGPT in their classes.
But not every technology leader is embracing the new chatbots with the same fervor. The latest to express doubts is Vint Cerf, Google’s chief internet evangelist—and a man regarded as the “father of the internet.”
Cerf had one piece of advice for investors who want in on the chatbot hype—whether that’s with ChatGPT, Bart or any other application: to not get swayed just because it’s “cool.”
Cerf tried dabbling with the tool to create a biography about himself, and it failed—the final outcome was full of factual inaccuracies, according to CNBC. He also tried asking the tool to add an emoji at the end of the sentence, which it didn’t do and apologized for.
“Everybody’s talking about ChatGPT or Google’s version of that and we know it doesn’t always work the way we would like it to,” Cerf told CNBC on Monday at the Celesta Capital’s TechSurge conference in Mountain View, Calif., referring to Bard, which was unveiled last week. “There’s an ethical issue here that I hope some of you will consider.”
“If you think, Man, I can sell this to investors because it’s a hot topic and everyone will throw money at me, don’t do that,” Cerf said. He urged investors to “be thoughtful” about investing in the technology.
Chatbots are far away from self-awareness, he added, saying that they can not differentiate between eloquent and accurate responses.
“Depending on the application, a not-very-good fiction story is one thing. Giving advice to somebody…can have medical consequences. Figuring out how to minimize the worst-case potential is very important,” Cerf said about how the tool may be put into effect.
Cerf isn’t the only one sounding alarms about how generative A.I. technology needs to be thought about with caution. Other tech leaders including Apple cofounder Steve Wozniak have said that platforms like ChatGPT have the potential to make “horrible mistakes.” And Google’s search engine chief acknowledged earlier this week that chatbot tools could experience “hallucination,” producing “convincing but completely fictitious answers.”
ChatGPT’s parent company, OpenAI, declined Fortune’s request for comment on this story.
Update, February 15, 2023: This article was updated with the name of the conference where Vint Cerf spoke.
