The Race to Deploy Doctor-Free AI: Inside the Startups Exploiting America's Regulatory Patchwork

The era of AI as a physician's helpful assistant may be ending before it truly began. While tech giants like Anthropic position their healthcare AI as supportive "copilots," three ambitious startups are pursuing a far more disruptive vision: artificial intelligence systems that diagnose conditions and prescribe medications entirely without physician oversight.

Doctronic, Certuma, and Ada Health represent the vanguard of what could become healthcare's most contentious transformation—the replacement of human clinical judgment with algorithmic decision-making. Their strategies reveal how companies are exploiting regulatory gaps at state and international levels to deploy autonomous medical AI years before federal frameworks catch up.

Utah's Regulatory Sandbox Becomes Testing Ground

Doctronic, a New York-based startup that raised $40 million in Series B funding this March, is already operating under Utah's AI regulatory sandbox—a non-enforcement agreement that allows the company to autonomously approve prescription renewals for chronic conditions without physician sign-off.

The six-month pilot data presents a mixed picture. Doctronic's AI independently approved prescription renewals 72% of the time across 192 drugs treating hypertension, diabetes, and depression. When the system escalated cases to human physicians (28% of the time), independent reviewers agreed the escalation was appropriate in only 69% of cases. More concerning: physicians disagreed entirely with AI renewal decisions in 3% of cases, and researchers identified unresolved security vulnerabilities.

Despite these concerns, Utah's Office of AI Policy reported no serious safety incidents as of publication. The Utah Medical Licensing Board and American Medical Association have raised accountability questions, but the sandbox continues operating—demonstrating how state-level agreements can circumvent traditional medical oversight structures.

The FDA Approval Gambit vs. International Fast-Tracking

While Doctronic exploits state-level regulatory flexibility, Austin-based Certuma is pursuing the more arduous FDA De Novo pathway for autonomous diagnostic approval. Founded by Martin Varsavsky and backed by $10 million in seed funding from Joe Lonsdale's 8VC, Certuma targets 25 common conditions including UTIs, strep throat, and sinusitis.

The company's long-term vision extends beyond diagnosis to autonomous robotic booths capable of imaging and blood draws—a fully automated clinical encounter. Rather than wait for U.S. approval, Certuma is already operational in Argentina, where deregulation-friendly policies allow its diagnosis-and-prescribing chatbot to treat patients today.

This international arbitrage strategy highlights a critical reality: no autonomous AI prescribing service has received FDA clearance through any pathway as of mid-2026, despite the agency clearing over 1,000 AI/ML-enabled medical devices. The only fully autonomous AI diagnostic system ever authorized remains LumineticsCore's diabetic retinopathy detector, approved in 2018.

Symptom Checkers Scale While Regulations Lag

Berlin-based Ada Health represents a third approach: scaling globally before regulatory frameworks solidify. With $242 million in total funding and 13 million users worldwide, Ada has completed 32 million symptom assessments. A peer-reviewed BMJ study ranked Ada highest among eight symptom-checker apps on coverage, accuracy, and safety.

Ada's symptom-triage model occupies a regulatory gray zone—providing clinical guidance without explicitly diagnosing or prescribing. This positioning allows rapid international expansion while sidestepping the approval requirements that constrain Certuma and the state-level controversies facing Doctronic.

The Regulatory Fragmentation Problem

The divergent strategies of these three companies expose healthcare AI's central challenge: regulatory incoherence. By mid-2026, 37 states had passed or introduced AI-in-healthcare legislation, creating a contradictory patchwork. Oregon and Delaware explicitly bar "nonhuman entities" from using licensed clinical titles, while California, Texas, and Nebraska require licensed human clinicians for final medical necessity determinations. Utah's sandbox model directly conflicts with these restrictions.

This fragmentation has real workforce consequences. Revere Health eliminated approximately 200 jobs—7% of its workforce—in medical coding, billing, and documentation roles due to AI implementation. A Sermo poll found 58% of physicians believe AI will diminish their role or make doctors obsolete.

The fundamental question remains unresolved: when an autonomous AI system makes a clinical error, who bears liability? Until malpractice frameworks, credentialing standards, and federal regulatory pathways catch up to technological deployment, patients will navigate a healthcare system where the rules governing their care depend on which state they live in—and whether their provider chose to exploit regulatory gaps or wait for comprehensive oversight.

Sources

1. Here Are 3 AI Companies Betting They Can Replace Doctors — Forbes