Though Steve Wozniak considers ChatGPT to be “fairly spectacular” and “helpful,” he additionally mentioned it will possibly royally screw up resulting from its lack of “humanness.”

“I believe it should be helpful to people as all pc know-how [is] … AI is on the excessive finish of it, the place it follows procedures to be taught issues,” the Apple co-founder mentioned on CNBC’s “Squawk Field” (see video under).

“It is fairly spectacular,” he continued. “However the hassle is … it will possibly make horrible errors.”

I want Wozniak would have gone into extra element of what these horrible errors may appear to be, however he merely in contrast his issues about ChatGPT with these of self driving automobiles. “It’s going to by no means be equal … Good lord, have a look at plenty of AI examples the place you drive a automobile and it makes selections the dumbest driver on this planet would not make.”

From CNBC:

Wozniak pointed to self-driving automobiles as a technological improvement with comparable issues, noting that synthetic intelligence cannot at present change human drivers. “It is such as you’re driving a automobile, and you already know what different automobiles could be about to do proper now, as a result of you already know people,” he mentioned. …

ChatGPT’s know-how can actually assist people — by explaining coding languages or setting up a body to your résumé, for instance — even when it does not but know methods to convey “humanness” or “feelings and emotions about topics,” Wozniak mentioned. …

Wozniak is not the one tech billionaire cautious of these penalties.

ChatGPT and its dad or mum firm, OpenAI, have “gorgeous” web sites — however they’re certain to be corrupted by misinformation as they internalize extra data throughout the web, serial entrepreneur and investor Mark Cuban informed comic Jon Stewart’s “The Downside with Jon Stewart” podcast in December.

“Twitter and Fb, to an extent, are democratic inside the filters that an Elon [Musk] or [Mark] Zuckerberg or whoever else places [on them],” Cuban mentioned. “As soon as these items begin taking up a lifetime of their very own … the machine itself could have an affect, and it will likely be tough for us to outline why and the way the machine makes the selections it makes, and who controls the machine.”

Wozniak’s issues over ChatGPT’s “humanness” remind me of what students had been speaking about in Jennifer Sandlin’s publish on Monday, the place one professor mentioned the app “fails spectacularly” when it’s requested for an “evaluation or critique” of, say, a novel. It could spit out any form of information you ask for and organize the information into articles and essays, however not less than for now, people have the higher hand in terms of true pondering.

Entrance web page thumbnail picture: CNBC (screengrab)