AI models could fuel "large-scale disinformation" as well as cyber-attacks. - OpenAI CEO

Must Read

OpenAI CEO Sam Altman told ABC News that he is worried AI models could fuel "large-scale disinformation" as well as cyber-attacks. 

He admitted that AI "is going to eliminate a lot of current jobs," but people "can make much better ones."

More:

Altman made the comments during an interview with Rebecca Jarvis, ABC News' chief business, technology, and economics correspondent.

He told Jarvis that when it comes to AI, "we've got to be careful here," adding, "I think people should be happy that we are a little bit scared of this."

When asked why he was scared, Altman said that if he wasn't, "you should either not trust me or be very unhappy that I'm in this job."

Despite the potential harms, he said AI could be "the greatest technology humanity has yet developed."

GPT-4:

The CEO discussed OpenAI's new GPT-4 large language model, acknowledging that it's "not perfect" but did score high on the bar exam and SAT math test and can write code in most programming languages.

Altman said the language model is a "tool that is very much in human control" since it waits for the user to prompt it with an input.

However, there will be those "who don't put some of the safety limits that we put on," he said, apparently referring to OpenAI.

"Society, I think, has a limited amount of time to figure out how to react to that, how to regulate that, how to handle it," he added.

A reasoning engine:

Altman also cautioned people about the models' "hallucinations," in which they "confidently state things as if they were facts that are entirely made up."

He said it's more correct to view the models as "reasoning engines," not as fact databases.

Facts are "not really what's special about them," he said, adding that "what we want them to do is something closer to the ability to reason, not to memorize."