October 24, 2025 brings key AI developments across medical practice, drug development, and professional governance. Today's updates show measurable clinical benefits, regulatory milestones, and institutional responses to AI adoption challenges.
AI scribes deliver dramatic burnout reduction in two health systems
China achieves first AI-designed drug formulation to complete Phase III
AMA launches center to embed physicians in AI development
Harvard reveals simple tactic to reduce AI bias
Health systems report 20% fewer unnecessary ER visits with AI triage
Marketing agency runs campaign directing patients from AI chatbots to real doctors
Have suggestions? Reply to this email.
AI Scribes Cut Burnout, Free Up Face-Time — Big Drops Seen in Two Systems
Two US health systems ran pilots of AI ambient documentation that listens to visits and drafts notes. Mass General Brigham saw burnout fall from 52.6% to 30.7% after 84 days — a 22-point drop. Source At Emory, clinicians reporting that documentation positively affected well-being rose from 1.6% to 32.3% after 60 days. Source
Usage rates were strong: 50% of Mass General Brigham clinicians used AI scribes for at least half their visits. Some clinicians reclaimed one hour per day and reported more eye contact with patients. However, some noted bulkier notes and extra editing time. Response rates were modest at 22% and 11% respectively, suggesting need to monitor adoption patterns before scaling.
China's First AI-Designed Formulation Reaches Phase III — METiS's MTS-004
METiS BioTech announced that MTS-004 is the first drug in China with an AI-designed formulation to complete Phase III development. Source This milestone shows AI can scale from lab to late-stage clinical development.
The achievement signals potential for AI to speed time to market and cut formulation costs. It also demonstrates commercial viability for AI-enabled pharmaceutical assets in major markets. Firms should track regulatory filings and next steps for commercial timing insights.
AMA Launches Center for Digital Health and AI — Puts Physicians in Charge of AI Design
The American Medical Association launched a Center for Digital Health and AI to embed physicians throughout AI tool design and deployment. AMA CEO John Whyte said physicians must be "embedded throughout the lifecycle of technology development and deployment." Source
The move complements recent sector efforts including Joint Commission guidance on responsible AI and WHO ethics guidelines. Buyers can now expect stronger clinical input in AI products, which should reduce rollout risk and ease clinician adoption. The Center raises the bar for clinician-led AI governance.
Make AI Blind to Biasing Data — A Simple Fix That Works
Harvard Business Review describes one simple move to cut bias in AI: deny the model the data you think causes bias. Source Run "blind" models without suspect variables, then compare outcomes to models that had them.
This low-cost tactic reduces legal and reputational risk while providing clear evidence on whether a feature drives unfair outcomes. It helps leaders decide what to remove or audit. Use blind tests with fairness metrics and monitoring as a practical first step before complex fixes.
AI Cuts Docs and Directs Care — Geisinger Sees 20% Fewer Unneeded ER Visits
Health system leaders at HLTH 2025 shared early ROI from AI pilots. Risant Health's Intelligent Triage cut unnecessary ER encounters at Geisinger by about 20% and reduced unnecessary specialty referrals by 7% over the last year. Source
Geisinger reports clinicians save roughly one hour per day with ambient notetaking tech. Houston Methodist saw mortality drops after adopting AI for monitoring and plans to shift 20% of specialty care to virtual settings in three years. These results show AI lowering costs while freeing clinician time.
Havas Lynx Hijacks "Chat GP" to Stop AI Medical Misinformation
Havas Lynx ran a campaign that "hijacked the AI conversation" to warn people not to trust AI-only medical answers. The agency told viewers: talk to your GP for health advice you can trust. Source
The campaign framed AI chat as a place where misinformation can spread. Messaging was simple: use AI for info, but see your GP for care. The content reached both patients and clinicians through professional and social channels. This provides a clear model for firms wanting to fight AI health misinformation while protecting trust in clinical care.
Sources
These developments show AI moving from experimental to operational across medical practice. The dramatic burnout reductions, successful Phase III completion, and institutional governance responses signal AI's growing integration into