Musk doubles down on artificial intelligence at Tesla amid scrutiny of autopilot

Mr. Musk, Tesla’s chief executive, said Tesla would build a robot in a human form that could perform repetitive tasks, with a prototype likely to be ready next year. It would draw on some of the technology Tesla has developed for vehicles, he said Thursday during an event in Palo Alto, Calif., around artificial intelligence, or technology designed to mimic the way humans think.

Tesla, which has delivered far fewer vehicles than more established rivals, has risen to become the world’s most valuable car maker as investors bet in part on the potential of its technology.

“Tesla is much more than an electric-car company,” Mr. Musk said.

The deployment of such robots could fundamentally change the economy, potentially alleviating labor shortages, he said. “In the future, physical work will be a choice,” Mr. Musk said, adding that long-term, such a robot could make it necessary to provide a universal basic income, or a stipend to people without strings attached.

Mr. Musk, who was accompanied on stage by a human dressed as a robot, often uses such settings to drum up interest in Tesla products, though his bold predictions haven’t always come to pass. Two years ago at an event about automation, he projected that more than a million Tesla vehicles would be able to operate without a driver by the middle of 2020, positioning the company to launch a robot taxi service.

That hasn’t happened, though the company has made progress with some of its software.

“Generalized self-driving is a hard problem, as it requires solving a large part of real-world AI. Didn’t expect it to be so hard, but the difficulty is obvious in retrospect,” Mr. Musk said on Twitter last month.

Artificial intelligence, or AI, is at the heart of Tesla’s efforts to develop more advanced driver-assistance features and, eventually, fully autonomous vehicles. Andrej Karpathy, senior director of artificial intelligence at Tesla, and other company engineers delivered a highly technical peek under the hood of the company’s driver-assistance system.

Tesla’s Autopilot relies on sensors to help drivers with steering, maintaining a safe distance from other vehicles on the highway and other tasks.

The system has come under a microscope in recent days. The National Highway Traffic Safety Administration, which regulates auto safety, this week said that it was investigating Autopilot in the wake of 11 crashes since January 2018 involving Teslas at scenes to which emergency vehicles had responded. Such probes can but don’t always result in recalls.

Two U.S. senators have also asked the Federal Trade Commission to investigate whether Tesla has been deceptive in its marketing of Autopilot and an upgraded suite of driver-assistance features it calls Full Self-Driving.

Tesla hasn’t responded to requests for comment about the auto safety probe or the lawmakers’ letter and didn’t explicitly address the concerns at the event. Tesla has long said that Autopilot makes driving safer. Mr. Musk said on Twitter in April: “Tesla with Autopilot engaged now approaching 10 times lower chance of accident than average vehicle.”

Tesla has approached the challenge of developing autonomous vehicles differently than some other companies, increasingly relying on cameras to inform its vehicles about what is going on around them. Rival Waymo, the autonomous-driving arm of Google parent Alphabet Inc., collects data from many kinds of sensors. That mix, Waymo Co-CEO Dmitri Dolgov said in a Thursday blog post, “can reason more intelligently about the world.” Tesla didn’t respond to a request for comment on Waymo’s blog.

The companies’ strategies for getting autonomous vehicles into daily use also differ. Waymo has been using safety monitors as it validates its software, whereas Tesla has released driver-assistance features to the public as it works toward full autonomy.

Mr. Musk indicated that the company may introduce its pickup truck later than initially expected. He previously said that if Tesla got lucky, it would be able to deliver a few of the vehicles by the end of 2021. On Thursday, he said Tesla likely would debut new hardware in the so-called Cybertruck in about a year.

The Tesla event takes place amid booming demand at companies more broadly for AI expertise as businesses look to automate processes to become more efficient and to offer new services. Recruitment for AI positions has been rising across tech. Job postings this year are up roughly 41% from the year-ago period, according to CompTIA, the Computing Technology Industry Association.

Mr. Musk at the event encouraged people interested in working on AI applicable to real-world problems to join or consider joining Tesla.

Artificial intelligence has been a particular area of interest for Mr. Musk, who frequently talks about leveraging the technology to make cars autonomous. He also has warned of the risks posed by AI, describing it as an existential threat that could threaten jobs and even cause a war.

“It is the biggest risk that we face as a civilization,” he said of AI in 2017. He voiced concerns about AI at Thursday’s event, too.

Such warnings have sparked criticism. In a Twitter post last year, Facebook Inc.’s head of AI, Jerome Pesenti, questioned Mr. Musk’s understanding of the technology. “We are nowhere near matching human intelligence,” Mr. Pesenti said.

Mr. Musk replied: “Facebook sucks.”

The new Tesla Bot—to be roughly 5 feet 8 inches tall, weighing 125 pounds and able to lift 150 pounds—is “intended to be friendly of course and navigate through a world built for humans,” Mr. Musk said. Humans, he said, would be able to outrun it.

This story has been published from a wire agency feed without modifications to the text.