Elon Musk made some of his boldest claims yet concerning the future of artificial intelligence (AI) during an interview with CNBC’s Andrew Ross Sorkin.
During a wide-ranging interview, Musk responded to questions concerning recent lawsuits levied against some of the billionaire’s competitors in the AI space related to alleged copyright infringement.
Related: Elon Musk launches AI chatbot ‘Grok,’ says it can outperform ChatGPT
“So, you think it’s a lie,” Sorkin asked Musk during the interview, “when OpenAI says that… none of these guys say that they’re training on copyrighted data.”
Musk’s response, “Yeah, that’s a lie.”
Elon Musk’s digital god
It’s unclear what Sorkin meant by his query, as OpenAI has testified in court to the effect that it does train models on copyrighted material. It is the company’s stance, however, that doing so constitutes “fair use” under United States law.
Under further prodding from Sorkin, Musk dismissed the efficacy of the lawsuits by claiming that a “digital god” would make the copyright lawsuits irrelevant:
“I don’t know, except to say that by the time these lawsuits are decided, we’ll have digital god. So, you can ask digital god at that point. Um. These lawsuits won’t be decided on a timeframe that’s relevant.”
Artificial general intelligence
Assuming that Musk is referring to a digital god appearing within the next three to five years, the average time it takes for a class action lawsuit to conclude, this indicates that the mogul believes that such an entity is not only imminent but will have powers far beyond human control.
Musk has, in the past, claimed that Google co-founder Larry Page wanted to build a “digital god,” and he’s told both Meta AI boss Yann LeCun and United Kingdom Prime Minister Rishi Sunak that “our digital god is a .CSV file.”
Our digital god will be in the form of a csv file
— Elon Musk (@elonmusk) November 25, 2023
It appears Musk is referencing the idea of a superhuman intelligence, usually referred to as an “artificial general intelligence,” or AGI. While there’s no standard accepted definition of an AGI, it stands to reason that an entity capable of rendering the U.S. court system irrelevant in a matter of three to five years would sufficiently qualify as an AGI construct. And, arguably, such an entity would be well beyond human control.
Musk previously predicted that AGI would arrive before 2030, an estimate many industry experts disputed as overly optimistic.
New York University professor and best-selling author of Rebooting AI Gary Marcus and a host of other AI experts challenged Musk to a $500,000 bet on the matter, with the proceeds to go to charity, with their position being that AI would not arrive by 2030. Musk has, so far, declined to respond to the offer.
Related: ChatGPT’s first year marked by existential fear, lawsuits and boardroom drama