This analysis synthesizes sources published the week ending June 13, 2025. Editorial analysis by the PhysEmp Editorial Team.
The American Medical Association, the American College of Physicians, and bipartisan congressional sponsors all moved this week to draw firmer lines between physician judgment and artificial intelligence. For readers of Physician Compensation & Demand, that shift matters: AI governance language is already being used as a negotiating point in employment contracts, not just a compliance checkbox.
The AMA created a dedicated AI center and put it under the leadership of the Digital Medicine Society founder; the ACP released a formal position statement; Representative Landsman introduced bipartisan legislation to block AI from making autonomous healthcare decisions. Taken together, these moves give physicians institutional backing to demand contract language that preserves clinical authority. Employers who push back risk regulatory exposure and a harder time recruiting.
The Liability Gap Driving Contract Language
Several reports this week flagged a specific malpractice worry: ambient AI scribes that draft notes clinicians may sign without reading carefully. The legal question—can a doctor be held liable for AI-generated text they attested to but didn’t write—hasn’t been settled. That gap creates asymmetric risk when employment contracts require use of AI tools without spelling out who takes responsibility for mistakes.
Most contract templates don’t assign responsibility if AI-generated documentation contains errors that lead to harm. When employers make ambient AI scribes mandatory, they shift risk to employed physicians without extra pay or protective clauses. The ACP’s insistence that physicians keep “meaningful control” over AI is now a citation contract attorneys can use to argue for explicit liability carve-outs.
Physicians about to sign contracts with mandatory AI scribe clauses should insist on indemnification for AI-generated documentation errors, explicit time built into schedules for human review, and clear definitions of what counts as adequate attestation before agreeing to those terms.
Supervisory Authority as Compensation Component
The AMA’s five-responsibility framework—things AI can’t replace—creates a concrete argument during pay talks. If judgment, ethics, and the patient relationship are nontransferable duties, then contracts that let AI encroach on those duties without negotiation shrink the physician’s role without adjusting pay.
That matters because many productivity-based pay models assume AI will boost throughput. Employers who use ambient AI to speed visits capture the efficiency gains while physicians still carry attestation duties and legal risk. Physicians negotiating RVU-based deals can point to the emerging consensus and argue that any AI-enabled productivity lift should be shared with the clinician who signs the note.
The Recruitment Angle
Organizations hiring now face a clear choice: put protective AI governance language into offer letters, or lose candidates to places that do. Early adopters who bake AMA and ACP language into their contract templates will stand out to physicians who know the liability landscape. Recruiters should expect candidates to ask specific questions about AI governance; programs that can’t answer will look dated.
Legislative Backstop Creates Negotiating Floor
Landsman’s bill is not law, but it signals congressional intent to stop AI from making autonomous clinical choices. That gives physicians a bargaining chip: contracts that hand decision-making authority to AI could clash with future rules, creating renegotiation headaches and compliance costs for both sides.
Physicians signing multi-year deals with large signing bonuses should factor AI governance into their risk assessment. A three-year contract with a $75,000 signing bonus and mandatory AI scribe use looks very different if there’s no indemnification or review time spelled out.
What Employers Should Anticipate
Expect savvy candidates to arrive with specific requests: define AI as a tool, not a decision-maker; spell out liability for AI-created notes; build time into schedules for reviewing AI outputs; and include opt-outs when AI threatens care quality. Treating those asks as fringe demands will misread the market.
Organizations that add AI governance to their standard templates now will avoid long, individualized negotiations later and send a signal that they respect physician authority. Compensation goes beyond base pay: productivity incentives tied to AI need a risk-adjusted view. If bonuses depend on AI tools that increase clinicians’ legal exposure, the nominal numbers overstate the real value. Physicians should test whether those bonuses compensate for the attestation burden and negotiate when they don’t.
There’s also a potential market effect. Employers who refuse protective language may begin to attract clinicians who haven’t tracked these risks, while those who adopt protections pull in talent that has. Over time that could shift where experienced physicians choose to work, and who they choose to work with.
Picture a hiring packet, a late-night EHR alert, and a checkbox nobody read. That’s where this starts to feel real: a clause in a contract, not an abstract policy.
Sources
The Hidden Malpractice Dangers Hidden in AI Scribe Files – Medical Economics
Before AI Scribes Become Mandatory, Medicine Should Look at the Black Box – Becker’s Hospital Review
The Hidden Liability of Ambient AI – Becker’s ASC Review
Ethics and AI use in medical practice: American College of Physicians releases position statement – Dermatology Times
Physicians have 5 responsibilities that AI will not replace AMA says – Medical Economics
Landsman Introduces Bipartisan Legislation to Keep AI From Making Health Care Decisions – Office of Representative Greg Landsman
Digital Medicine Society founder to lead AMA’s new AI center – Becker’s Hospital Review
AI is already here — now healthcare has to figure out how to govern it – HC Innovation Group
