AMA Survey: 81% of doctors now use AI — so why are workers' comp clinics still stuck in 2015?

The number that should stop you

The AMA's 2026 Physician Survey on Augmented Intelligence reports that 81% of physicians now use AI in their professional practice — more than double the 38% reported in 2023. That isn't a curve. It's a wave.

If your occ med clinic or workers' comp health system hasn't started, you're not behind the curve — you're under it.

What the survey found

The AMA surveyed 1,692 physicians across specialties and settings in early 2026. Physicians now average 2.3 active AI use cases (up from 1.1 in 2023). Nearly 40% already use AI for medical-research summaries — a 26-point jump in one year. Documentation tools (chart summaries, discharge instructions, billing) are the next frontier, with more than half of physicians expecting to adopt them within the year.

Sentiment is overwhelmingly positive

75% say AI provides a clinical advantage. 70% see AI automating tasks that contribute to burnout. 73% expect it to reduce administrative load.

But trust depends on guardrails

88% say validation of safety and efficacy is critical. 86% require data-privacy assurances. 31% rank clear liability frameworks as the single most important regulatory priority. Physicians aren't asking whether AI is coming; they're asking who owns the risk.

Training is wanted, scarce

92% want more AI training. Only 11% of those who've received training call it extensive.

Why this matters more in occ med and workers' comp

Workers' compensation healthcare operates in a pressure cooker the rest of clinical practice doesn't face the same way. Multi-party stakeholders, heavier documentation, tighter turnaround SLAs, thinner margins. The exact pain points the AMA's survey highlights are where the value lands hardest.

Documentation: the biggest quick win

Documentation is ground zero. In occ med it maps directly to faster RTW narratives, more consistent injury documentation, and streamlined utilisation review responses. The clinical judgement doesn't change. The 45 minutes of documentation that follows does.

Diagnostic support and risk prediction

74% of surveyed physicians expect AI to improve diagnostic ability. In workers' comp, where early accurate diagnosis directly drives claim duration and cost, that's a major lever — especially for musculoskeletal injuries and predictive flags on claims trending toward chronicity.

Burnout in a workforce you can't afford to lose

Occ med already has recruitment challenges. Losing experienced clinicians to burnout means losing institutional knowledge about employer relationships, hazard patterns, and regulatory nuance. AI that takes the documentation grind off providers' plates doesn't just make the day more bearable — it makes the business sustainable.

The privacy question hits differently here

The AMA survey found patient privacy is the only dimension where more physicians expect AI to cause harm than help. In workers' comp, that concern is sharper: records flow between more parties, HIPAA intersects with state-specific workers' comp privacy rules, and data mishandling carries real stakes. Any AI deployment needs rock-solid governance from day one — see our piece on why most AI projects fail on data quality.

The bottom line

Physician adoption has crossed the tipping point. Patients expect it. Organisations that implement it thoughtfully will outperform those that don't. For occ med leaders, the question isn't whether to adopt AI — it's whether you lead the adoption curve in your market or react to competitors who did. Talk to Alana Health.