Apple co-founder Steve Wozniak has lambasted Tesla’s AI-powered autopilot function, calling it “a study of AI gone wrong" that will try "to kill you every chance it can” in a new interview with CNN This Morning.
The show’s hosts asked Wozniak to speak about his relationship with Tesla, SpaceX, and Twitter CEO Elon Musk, to which Wozniak replied “I never actually met him and spoke with him. I admire some of the things he’s done for the world, changing us towards electric cars.”
But Wozniak’s own experience as a two-time Tesla customer appears to have put the Silicon Valley magnate off of Tesla’s self-driving promises. In the interview, Wozniak blamed Musk for making promises that, Wozniak felt, were false.
“He basically got a lot of money from myself for cars,” Wozniak said, before adding “I believed things he said—a car would drive itself across the country by the end of 2016. I had to upgrade to that model, you know, $50,000… and it wouldn’t do anything.”
“I could tell it wouldn’t make it across the country” he summarized. He then admitted to making the same expensive mistake a year later after Musk advertised “a new one with eight cameras [that’ll] make it across the country by the end of 2018.”
Wozniak says Musk’s marketing spiel is “not even close to reality and boy if you want a study of AI gone wrong and taking a lot of claims and trying to kill you every chance it can, get a Tesla.”
This is not the first time that Wozniak has used hyperbolic language to express his Tesla regret. Back in February, he told Fortune magazine that Tesla has “robbed my family of so much money” after he purchased the two electric vehicles.
Great minds think alike
Despite Wozniak’s sharp criticisms of Tesla’s AI functionality, both he and Musk share a common cause: both are concerned by the rate of development of artificial intelligence systems.
In March this year, the pair were leading signatories to an online petition calling for ChatGPT developer OpenAI to halt development for six months on any system more clever than GPT-4, the latest iteration of the intelligent, multi-talented chatbot.
Their concerns are not entirely unwarranted. Many are alarmed by the things GPT-4 is already capable of, including exploiting smart contracts, rudimentarily planning its own escape and falsely accusing people of committing sexual assault—hallucinating a Washington Post article as a citation to substantiate the claim.
The OpenAI team appears to have heeded their warnings. Last month, developer Logan Kilpatrick assured his followers on Twitter that work has not commenced on GPT-5 and will not “for some time.”