Will we ever trust robots?

DAVID VINTINER

“A lot of companies that do that kind of stuff end up doing it in a way which is kind of shitty for the people who are being employed,” Hashme told me. Such companies often outsource important HR activities to untrustworthy partners abroad or lose workers’ trust through bad incentive programs, he said, adding: “With a more experienced and closely managed team, and a lot more transparency around the entire system, I expect we’ll be able to do a much better job.”

It’s worth disclosing the nature of Hashme’s departure from Scale AI, where he was hired in 2017 as its 14th employee. In May 2019, according to court documents, Scale noticed that someone had repeatedly withdrawn unauthorized payments of $140 and transferred them to multiple PayPal accounts. The company contacted the FBI. Over the course of five months, approximately $56,000 was taken from the company. An investigation revealed that Hashme, then 26, was behind the withdrawals, and in October of that year, he pleaded guilty to one count of wire fraud. Ahead of his sentencing, Alexandr Wang, the now-billionaire founder and CEO of Scale AI, wrote a letter to the judge in support of Hashme, as did 13 other current or former Scale employees. “I believe Shariq is genuinely remorseful for his crime, and I have no reason to believe he will ever do something like this again,” Wang wrote, and he said the company would not have wanted the wrongdoer prosecuted if it had known it was Hashme. 

Hashme lost his job, his stock options, and Scale’s sponsorship of his green card application. Scale offered him a $10,000 severance payment before leaving, which he declined to accept, according to Wang’s letter. Hashme paid the money back in 2019, and in February 2020, he was sentenced to three months in federal prison, which he served. Wang is now a primary investor in Prosper Robotics, alongside Ben Mann (cofounder of Anthropic), Simon Last (cofounder of Notion), and Debo Olaosebikan (cofounder and CEO of Kepler Computing). 

“I had a major lapse in judgment when I was younger. I was facing some personal challenges and stole from my employer. The consequences and the realization of what I’d done came as a shock, and led to a lot of soul-searching,” Hashme wrote in an email in response to questions about the crime. At Prosper, he wrote, “we’re taking trustworthiness as our highest aspiration.”

There are some real upsides to being able to control robots remotely, but the idea of large-scale robotic teleoperation by overseas workers, even if it takes years for it to be effective, would be nothing short of a seismic shift for labor. It would present the possibility that even highly localized physical work that we perceive as immune to moving offshore—cleaning hotel rooms or caring for hospital patients—might someday be conducted by workers abroad. It also seems antithetical to the very idea of a trustworthy robot, since the machine’s effectiveness would be inextricably tied to a faceless worker in another country, most likely receiving paltry wages. 

Hashme has spoken about using a portion of Prosper’s profits to make direct payments to people whose jobs have been affected or replaced by Alfies, but he doesn’t have specifics on how that would work. He’s also still thinking through issues related to who or what Prosper’s customers should be trusting when they allow its robot into their home. 

“We don’t want you to have to place as much trust in the company or the people the company hires,” he says. “We’d rather you place trust in the device, and the device is the robot, and the robot is making sure the company doesn’t do something they’re not supposed to do.” 

He admits that the first version of Alfie will likely not live up to his highest aspirations, but he remains steadfast that the robot can be of service to society and to people, if only they can trust him.