05 Sci-Tech

On January 8th, OpenAI introduced ChatGPT for Healthcare, a generative AI (GAI) platform designed to be embedded within medical systems platforms and daily workflows. This technology suite is advertised as a solution to clinicians overburdened by administrative work through offloading cognitively taxing tasks, including the choice of diagnostic tests, supporting differential diagnosis, treatment planning, documenting session notes, creating aftercare plans for patients, and generating referral notes and discharge summaries for external providers. In other words, GAI is being implemented at every level of patient care. According to the American Medical Association’s report from their summit on AI, “disruption” of the status quo in healthcare delivery due to GAI technologies “seems inevitable.”

But why does it seem inevitable? An evidenced-based approach to evaluating new technologies would call for careful consideration of benefits and risks for technology implementation on individual use cases — not a rapid systems overhaul. Here, we must recognize that GAI technologies are products — and these products are being actively promoted to healthcare industries and healthcare professionals across the medical space, including in mental health care. Rather than investing billions of dollars into curtailing a failing system of private medical care — which has led to widespread clinician burnout and poor client outcomes — Silicon Valley companies have begun attempting to mud over these fault lines with a quick-drying GAI compound. Even the most well-meaning and justice-oriented clinician is not immune to the tidal wave of billion-dollar marketing strategies bent on creating the illusion of inevitability.

The Pentagon rarely labels an American technology company a “supply chain risk.” The designation is typically reserved for firms tied to foreign adversaries or companies that could expose sensitive government systems to compromise.

But in late February, the Trump administration applied that label to one of the most prominent artificial intelligence developers in the United States.

On Monday, Anthropic, the company behind the Claude AI system, turned up the heat on the fight by filing a federal lawsuit against the Pentagon and several government agencies after the administration ordered agencies to stop using its technology across the federal system.

“Anthropic sued the Defense Department and other federal agencies on Monday over the Trump administration’s move to designate it a supply chain risk and eliminate its use across the government,” the report explains. “The company said the effort was ‘unprecedented and unlawful.’”

Calls for governments to push “pro-worker AI” sound appealing. The idea is simple: If policymakers deftly guide how the technology develops, they can make sure it helps workers instead of replacing them. What’s not to like?

Here’s your trouble: Technology almost never works that neatly. Its effects on jobs are usually messy, unpredictable, and shaped by millions of decisions from businesses and entrepreneurs—not by a policy plan designed in Washington.

That’s a core point in a recent critique by economist Joshua Gans of a proposal from Daron Acemoglu, David Autor, and Simon Johnson to steer AI toward worker-friendly uses. Gans says the idea runs into a basic contradiction. The proposal defines “pro-worker” technology as something that makes human capabilities and expertise more valuable. But those things are valuable partly because not everyone has them. If a new technology spreads skills more widely, it may help more workers overall—while at the same time reducing the pay advantage of those who once had rare skills.