UnitedHealthcare, the largest insurance company in the United States, leveraged artificial intelligence to slash $100 million in costs — at the expense of 300 nurse practitioners’ jobs and maybe even the chief executive’s life, suggested a healthcare consultant yesterday.

Clayton Gillett, managing partner at Info-Tech, a healthcare consultancy firm, spoke to an audience of chief information officers and other healthcare executives at Healthcare CIOConnect Bermuda, convened by KPMG.

He said the move, which saw AI systems reviewing medical records and suggesting approvals for procedures, has ignited debate over the human cost of technological progress.

Mr Gillett pointed out the dramatic shift: “Their board said, spend all you want on AI, as long as you give me $100 million in return on investment this year. And then they said, If you can't do that, you shouldn't be our CIO.”

The result?

“Three hundred nurse practitioners were let go,” Mr Gillett said. “Three hundred people moved, changed positions and that CIO said he saved $100 million.”

But the savings came with a catch. “At the same time, their rejection rate went from 11 per cent to 22 per cent and the CEO [Brian Thompson] got shot, and most people in the US did not care,” Mr Gillett noted.

Brian Thompson was the CEO of UnitedHealthcare until he was shot and killed in Manhattan last December, allegedly by Luigi Mangione, who faces federal murder and stalking charges, along with state charges. The motive for the killing is believed to be related to anger over US health insurance companies.

Mr Gillett’s presentation also noted other eye-catching examples of corporate AI use with mixed results, such as Microsoft’s $24 billion AI investment, which increased its electricity bill by $200 million.

He highlighted a Colorado children’s hospital CIO who retrained staff displaced by AI tools. “In the first year, she replaced 20 people with IT tools they already had. She got 20 staff, which she brought into IT,” he said.

Under her agreement with the company, the staff’s retraining was funded. Others trained to be nurses, and the CIO moved on to the next money-saving project.

“They replaced their entire telephone system next with Google phones, and made them soft phones. They did all that with a million dollars. And they also made another round and started to do diagnostic-type AI with machine learning to improve early detection.”

The UnitedHealthcare case, however, Mr Gillett argued, was a cautionary tale: “You need to have governance. You need to understand what you’re shooting for. You need to hold yourself accountable for doing that.”