The man selling cheap drugs just predicted AI's worst-case healthcare scenario — and it's already happening.
The Summary
- Mark Cuban says AI won't fix US healthcare because insurers will weaponize it against doctors and patients, creating an AI-versus-AI arms race where "conglomerates will have multiple adversarial agents doing all they can to delay and deny"
- He claims 25% of physician time already goes to fighting insurance friction, and AI will automate that friction at scale
- Hospitals are paying up to 10% of revenue to Revenue Cycle Management firms whose AI agents fight insurer AI agents over claim denials
The Signal
Cuban is describing the nightmare version of the agent economy: not human-versus-machine, but machine-versus-machine with humans caught in the crossfire. He was responding to Marc Andreessen's claim that "AI is already a better doctor than 99.99% of human doctors" by pointing out that diagnostic capability means nothing if the system itself is adversarial by design.
The economics are perverse. Insurers make money by denying claims. Doctors make money by providing care. Right now, that conflict burns human hours on both sides. Prior authorizations. Claim appeals. Endless phone trees. Cuban says doctors already spend a quarter of their time on this friction. AI doesn't eliminate the friction. It automates it. Both sides.
"For every future agent we give AI doctors to deal with this friction, the conglomerates will have multiple adversarial agents doing all they can to delay and deny."
Here's what that looks like in practice:
- Doctor's AI agent submits a claim for a scan with supporting diagnosis codes
- Insurer's AI agent instantly cross-references against policy fine print, finds a technicality, denies
- Hospital's Revenue Cycle Management AI (which takes 10% of revenue) auto-appeals with additional documentation
- Insurer's second-tier AI agent evaluates appeal, finds another reason to delay
- Loop continues until someone gives up or the patient pays out of pocket
Cuban says this is already happening. Insurers are using AI to "manipulate contracts and find ways to mislead." Hospitals are hiring firms to deploy counter-agents. It's an arms race where both sides are spending resources to fight each other instead of helping patients.
The math doesn't work for patients. If hospitals are paying 10% of revenue to firms that fight AI with AI, that cost gets passed through. If insurers are spending on adversarial agents to protect margins, premiums go up. The only winners are the companies selling AI agent infrastructure to both sides. Everyone else is paying more to run in place.
Cuban has credibility here. He cofounded Cost Plus Drug Company in 2022 specifically to cut out pharmacy benefit managers, the middlemen who negotiate drug prices. He's built a business on eliminating intermediaries. Now he's watching AI create new ones, or at minimum, make the existing ones faster and harder to bypass.
The Implication
This is what happens when you deploy agents in adversarial systems without changing the incentives. Healthcare doesn't need better AI. It needs aligned AI. Right now, insurer AI is optimized for profit protection. Provider AI is optimized for claim recovery. Patient outcomes are a side effect, not the objective function.
Watch who's selling picks and shovels in this arms race. Revenue Cycle Management firms. Healthcare AI startups that pitch "claim defense" or "denial prediction." They're selling to both sides. If you're building in this space, ask whether you're solving the problem or automating the dysfunction. The distinction matters.