3 Red Flags That Your AI Is Working Against You: How to Avoid Cognitive Deskilling

Two thirds of physicians now use health AI in some form, up 78 percent from 2023. Adoption is no longer the question. The cost of speed is. That cost has recently been labeled the deskilling dilemma. It's what happens when the tools you adopted to sharpen your judgment quietly begin to replace it. You use AI more. You think through problems less. The output looks right, so you stop checking whether it is.

This week Dr. Sunjya Schweig, Dr. Lexi Gonzales, and Sunita walked the Ultralight community through the four foundational skills every modern clinician needs during a live AI Skills for Clinicians webinar.. The frameworks are durable. The conversation that surprised us most was about what happens after you have built them. Three red flags emerge in heavy AI users that no platform marketing will warn you about. Catching them early is the next skill.

OvationLab developed the AI in Clinical Practice skillset to streamline modern clinicians integrating AI into their daily practice with confidence and competence.

Here is a summary of the foundations we covered:

Friction mapping. Walk through your clinical week and name where the work feels harder than it should. Match each friction point to an AI tool or workflow change.

Iterative refinement. Never accept the first output. The first prompt gets you to 60 percent. The third prompt with context and examples gets you to 90.

Collaborative co-review. Bring AI into the visit on purpose. Share your screen with the patient. Turn the search into a shared one.

Ambient communication. Verbalize the clinical thinking you used to do silently so the scribe can capture it. The note gets richer. The patient hears your reasoning.

These work (until they don’t). Let’s dive into what to watch out for.

You stop reading the AI note carefully because it looks good. You sign off on a care plan you did not fully think through. You take the differential the LLM gave you without running your own. 

It’s automation bias. AI output is fluent. Fluent reads as correct. Your brain offloads the work it would normally do because the answer arrives faster than your skepticism. Learners can become deskilled, miscalibrated, or never skilled if they over rely on language models. As Dr. Sunjya put it during the webinar, “You just kind of accept what the AI gives you because you get so used to it.”

To fix this, add a friction step back to your workflow. Read every AI generated note out loud once before signing off. Run your own differential before consulting the AI. Treat the AI as a second opinion. If you cannot articulate why the AI was right, you have not done the thinking yet.

You read what the AI produced and feel like you understand it. A week later a patient asks why you chose that protocol and you cannot reconstruct the reasoning. You were touring the model's thinking instead of building your own.

Generated text feels familiar even when it is not yours. Your brain does not get the workout of constructing the argument because the argument arrived prepackaged. Over time your clinical instincts dull in the areas you have outsourced most. This shows up especially in nuanced functional and longevity workups, where the reasoning across labs, symptoms, and lifestyle history is exactly what your patient is paying you for.

You need to force yourself to teach what the AI gave you. If you cannot explain a protocol to a colleague or a patient in your own words, you have not learned it. Pick one area each week where you reason without AI assistance. Keep the cognitive muscle warm. The most experienced clinicians we know schedule unplugged thinking blocks on purpose.

You feel more tired than you did before AI, even though the documentation burden is lower. Every output needs a review. Every draft needs an edit. You have quietly become a full time editor of work you used to produce yourself.

Reviewing is cognitively different from creating, and more tiring than most people expect. Software engineers feel this when they go from writing code to reviewing AI generated code. Clinicians feel it when the day fills with AI drafted messages, notes, plans, and lab summaries that all need verification. Review fatigue is one of the most underdiscussed costs of heavy AI use.

Try to cluster your reviews. Set two or three review windows per day instead of bouncing into the chart every fifteen minutes. Build trust calibration with your tools. Tools that are reliably good get a faster review. Tools that are not get cut. Protect at least one block per day where you are creating instead of editing.

Ask yourself three questions:

A yes to any of those is a signal to adjust your workflow before the pattern compounds. Your job is to notice the shifts in the model and in yourself.

Catch the red flags early and your AI stack stays an amplifier instead of a substitute.

Reply and tell us which red flag you noticed in your own week. The best ideas in this newsletter come from clinicians doing the work.

Until next Friday, keep building the practice you imagined when you started.

— Sunita and Dr. G

Be a modern clinician with the help of Ultralight, the AI-native EHR built specifically for functional, integrative, and longevity medicine.

Patients are turning to Dr. Chatbot for primary medical advice over traditional clinicians. Driven by systemic friction like long appointment wait times and brief visit windows, consumers are increasingly using conversational LLMs to self-diagnose and vet treatments before ever speaking to a doctor. As these tools become a default first stop, the clinician's role is shifting rapidly toward validating and contextualizing information the patient has already surfaced. Worth assessing how your practice addresses self-directed AI diagnoses when patients arrive with pre-packaged chatbot protocols.

OpenEvidence introduces a hands-free voice interface for clinical decision support. Launched as a native speech-to-speech feature, the new OpenEvidence Voice Mode allows doctors on active rounds or moving between patient rooms to speak clinical questions aloud and hear concise verbal answers synthesized from peer-reviewed literature. The release marks a major push into enterprise medicine alongside a system-wide rollout at Cedars-Sinai that merges global medical literature with local patient data. Worth testing if you need a quick, evidence-backed way to cross-reference guidelines without looking at a screen.

Wolters Kluwer deploys a safety verification framework for clinical AI tools. Moving beyond generic software benchmarks, the new Wolters Kluwer Clinical AI Validation Framework gives health systems a rubric-based model to stress-test point-of-care generative utilities. The methodology relies on physician-led red-teaming and continuous workflow monitoring to catch omitted information, context loss, or unsupported medical claims before they reach a screen. Worth using to audit your own vendor's safety protocols and protect clinical intent.

Upcoming Conferences & Events

May 27–30, IFM Annual International Conference · San Diego, CA  · The largest gathering of functional medicine clinicians in the world. The clinical programming is worth the trip; the hallway conversations are worth twice that. Reply if you want to connect onsite. Ultralight team will be there!

Jun 9–11, Longevity Docs Cannes 2026 · Cannes, France · Invite-leaning room for clinicians at the frontier of longevity medicine. Worth it if you are designing the next iteration of your own practice.

Sep 24–26, Vibrant Longevity Summit · Austin, TX  · A clinical room of practitioners running lab-driven, longitudinal care. For anyone building a practice around diagnostics and biomarkers who wants peers who work the same way.

Oct 8–10, A4M Women's Health Summit · San Antonio, TX ·  The best clinical education on hormone, metabolic, and midlife women's health you will see this year. The room to be in if you are growing the perimenopause and menopause side of your practice.

Oct 21–24, NAMS Annual Meeting · San Diego, CA  · The single most practice-changing meeting of the year for midlife women's health. Your protocols will look different after this one.

Nov 5–8, Eudēmonia Summit · West Palm Beach, FL ·  One of the most talked-about longevity gatherings in the U.S. Experientials, hands-on demos, and the best place to try the emerging frameworks your patients will ask you about next year. Ovation and Ultralight team will be there!

Nov 5-7 — Private Physicians Alliance Annual Meeting · St. Petersburg, FL The gathering for independent, cash-pay, and concierge physicians navigating practice independence. Practical and peer-driven. Ultralight will be there!

Nov 8-11 — American College of Lifestyle Medicine Conference · Orlando, FL Lifestyle medicine's main annual event — evidence-based approaches to behavior change, chronic disease, and healthspan. Growing overlap with the longevity medicine community.

Dec 11–13, A4M Longevity Fest · Las Vegas, NV  · The biggest longevity event in the U.S. The room spans clinicians, industry, founders, and the people building next year's platforms, and the connections from this one tend to compound through the rest of your year. Ultralight will be there!

Know of an event we should add? Reply and tell us.