Six physician organizations publicly reject the claim that AI is "better informed" than doctors
October 6, 2026 · 5 items
Six physician organizations publicly reject the claim that AI is "better informed" than doctors
Medical Economics · October 1, 2026Practice operations
On Sept. 30 the AAFP, AAP, ACOG, ACP, ACS and AMA issued a joint statement saying claims that AI is inherently better informed than physicians — or that physicians can't be trusted to decide without consulting it first — "diminish physician expertise and risk undermining patients' trust." The statement came a day after the health secretary praised AI second opinions.
The groups flagged public assertions from federal leadership and AI industry representatives that framed AI consultation as essential and implied malpractice exposure for physicians who diagnose or prescribe without AI input.
AMA leaders challenged claims that AI-alone cognitive care is superior, citing the limits of simulation-based evidence and arguing for physician-governed integration within the care plan.
Medical Economics notes the liability picture is currently asymmetric: following an AI recommendation that deviates from the standard of care may increase exposure when outcomes are adverse — while low-risk operational uses (ambient documentation, summarization, billing) are widely accepted and decision support, ordering, renewals and messaging still require validation beyond vendor claims.
Why it matters for an independent practice: This is the authority argument your physician clients are about to be asked about by patients, and the specialty societies just handed them the language. For MMR: a short physician-authored post — "why I use AI for my notes and not for your diagnosis," citing the six-society statement — is credibility content that an AI summary can't manufacture, and it lands squarely on the low-risk/high-risk split the article draws.
Three California class actions now allege patients were recorded by an AI scribe without consent
Medical Economics · October 2, 2026Practice operations
A patient sued Sharp HealthCare in San Diego County Superior Court in late November 2025; a group of patients sued Sutter Health and MemorialCare in federal court in San Francisco in April. Both matters involve Abridge's ambient documentation tool and allege violations of California privacy law.
Health care attorney Tatiana Melnik, J.D. (Melnik Legal PLLC, Tampa) told Medical Economics that HIPAA applies to an AI tool "the moment" it can reach protected health information — whether it's being used for prior authorizations, scheduling or clinical documentation.
Melnik presented "Privacy and Security Legal Issues in the Age of Artificial Intelligence" on Sept. 28 at the MGMA 2026 Annual Conference in San Antonio, and sat for this interview at the conference.
Her first instruction to practices: review vendor contracts — existing and new — and understand exactly what consents the practice has already granted.
Why it matters for an independent practice: The consent script at the front desk is now a litigation surface, not a courtesy. For an independent practice running an ambient scribe, this week's action is to pull the vendor agreement and the intake consent form side by side and confirm the patient is actually being told a recording is happening — before someone else's class action sets the standard for yours.
Clinical AI is shifting physicians from doing tasks to overseeing agents that finish them
Healthcare IT News · October 1, 2026Practice operations
Dr. Talha Shaikh, chief of radiation oncology at New York Cancer & Blood Specialists/One Oncology, says most AI physicians use today is passive — it summarizes, drafts, answers — while agentic AI "can recognize that something is incomplete, take the next few steps, coordinate across different parts of the system and bring the physician back in when a clinical decision is actually needed."
Shaikh names the early use cases as the chases that eat clinic time: missing records, abnormal results needing follow-up, uncompleted referrals, prior authorizations, treatment prerequisites and patient messages. "The hard part isn't generating a good answer; the hard part is everything around it."
Joy Xu, UCLA medical student and lead author of the July 2026 npj Digital Medicine paper "AI agents in clinical practice: An evidence map," frames the change as execution-to-oversight — agents handle less cognitively dependent work like prior auth while physicians monitor outcomes.
Xu's condition on that: "True workflow transformation requires that problems be defined by physicians experiencing the burden, not by vendors offering systems." Dr. Nipa R. Shah of UF Jacksonville adds the practical ceiling — on Epic, you're dependent on what Epic can safely build and integrate.
Why it matters for an independent practice: This is the leverage argument stated in a form a practice can act on: list the five things your staff chases after every visit, then evaluate agents against that list rather than a vendor demo. Xu's "physicians define the problem, not vendors" rule is the buying posture that keeps an independent clinic from paying for someone else's workflow.
Apple is tightening macOS 'Full Disk Access' specifically because AI agents now use it
TechCrunch · October 2, 2026Buildable AI
Apple announced additional controls around the macOS "Full Disk Access" permission, warning that increasingly capable AI agents make broad access to a user's files, messages, mail and browsing history riskier.
The setting was originally designed so backup software could function properly — a permission granted once, years ago, for a mundane reason, now inherited by a very different class of software.
The announcement landed days after a journalist claimed Meta's Muse app on Mac read their private messages, a claim Meta disputed.
Why it matters for an independent practice: Every practice has at least one Mac with downloaded labs, scanned intake forms, billing exports and patient email sitting on disk. If an AI assistant was installed on it, somebody granted it a permission that reaches all of that — outside the EHR's access controls and almost certainly outside any BAA. The action is a ten-minute audit: System Settings → Privacy & Security → Full Disk Access, and revoke anything that isn't your backup tool.
OpenAI will invisibly watermark ChatGPT and Codex text — EU users only, API off by default
TechCrunch · October 5, 2026Buildable AI
OpenAI said Monday it will add an invisible watermark to text generated by ChatGPT and Codex in the European Union to comply with the EU AI Act, whose transparency rules took effect August 2 and require AI-generated content to be markable by other systems.
Rollout is over the coming weeks to eligible ChatGPT and Codex users on all plans, but only in the EU. Developers using the API anywhere in the world can switch it on for select models starting today — it is off by default, and OpenAI is not making watermarking a global default at launch.
The mark isn't a symbol: it subtly shapes the model's word choices into a pattern a detector can read but a reader can't see, and it survives copy-and-paste because it lives in the words themselves. OpenAI says it doesn't identify the user and saw no meaningful performance change.
TechCrunch notes OpenAI's own caveat that editing the text can make the marks harder to detect; the company published a technical report on the method, called textGrain.
Why it matters for an independent practice: Detectable provenance for AI text is now shipping infrastructure, not a theory — and for physician-authored content that is the whole ballgame. The safe posture for a practice blog, patient-education page or physician bio is unchanged and now more urgent: the doctor's judgment, examples and voice have to be genuinely in the draft, because "was this machine-written?" is becoming a question tooling can answer rather than guess at.
