To Succeed in Medical School, Resist AI Never-Skilling

- Reaching for an AI chatbot to build your differential saves minutes now but costs you something bigger later: the reps your brain needs to actually build lasting clinical judgment, a pattern researchers call "Never-Skilling"
- Feeling ready and being ready aren't the same thing. Studies show AI-assisted studying can boost confidence without touching accuracy
- Sketchy's visual method locks pattern recognition into memory on first exposure, so material doesn't have to be relearned every time boards, shelf exams, or an attending bring it back up
- Sketchy scales with the training: visual lessons and Step 1/Level 1 exam prep during preclinical years, Step 2/Level 2-focused case work once rotations start
- Sketchy subscriptions (Cases + Lesson Library + QBank) are 25% off with code BTS25
Training for an ultramarathon doesn't start the week before. Muscle gets built under load, through resistance, one day at a time.
Preparing to succeed in med school is no different. The clinical reasoning you'll need to do well on rounds, shelf exams, and the boards takes consistent effort. And that's a tough commitment to maintain for so many years. There's no finish line at the end of a semester, no long rest before the next one starts.
Because of this, I understand the appeal of cramming just before a high-stakes exam with an AI chatbot. It's like having a friend who's always there for you with an answer.
And sure, a friend might give you what you need right now. But sometimes you need a good coach who gives you what you need to become what you want to be. Because when it comes to medical school, putting in the effort yourself to reason through the correct differential pays off on exam day, and all the days in between.
What research says
In a 2026 study, medical, nursing, and emergency medicine students answered a survey about their true AI use. Forty percent named "reduced thinking" as their top fear, outranking concerns about whether the AI was even giving them correct information. Their fears, and this study, have been the topic of a lot of discussion here at Sketchy.
Researchers have started calling this phenomenon "Never-Skilling," and the idea is straightforward. When an AI chatbot carries the reasoning load early and often, the ability to reason itself is never given the chance to form in the first place.
The research is troubling. In a trial, novice medical students worked through diagnostic questions with AI explanations; some were correct, while others were deliberately wrong. The explanations didn't improve accuracy over none at all. And when the AI was wrong, students stayed just as confident. They were unable to tell sound reasoning from flawed (Teng et al., 2026).
In my experience, the reasoning you build yourself is what holds up when an attending puts you on the spot, or when you're alone with a shelf question and no support to fall back on. Lean on AI too early and the reasoning muscle doesn’t get worked; the scaffolding's gone the moment it actually counts.
Does this mean AI is off-limits in medical school?
Not at all. When you're stuck on a concept, especially the first time through, AI is genuinely great. It's excellent for comparing illness scripts, for clarifying something that your lecture or your textbook left unclear, sometimes even for generating practice questions to check whether a topic actually landed.
But you won't always have AI with you. On the shelf, on rounds, or in front of a patient, it's just you and what you’ve got in your head. And that's the real task: getting information into your brain reliably enough that you can retrieve it cold, without a computer to lean on.
That's where visual learning comes in. Sketchy's method locks a concept into memory the first time you see it, so when the moment comes to recall it on your own, it's there.
Students who put in the reps perform better under pressure
In my experience, successful students don’t necessarily spend more time studying than their peers. They simply train deliberately enough that recall starts becoming second nature.
Here's what that looks like in reality:
- They walk into an OSCE, shelf exam, or the boards with the reasoning already built
- They present on rounds without stalling to look something up first
- They still have room for sleep, for people outside medicine, for a life that isn't entirely coursework
For a lot of med students, Sketchy has already earned its place next to the white coat and the stethoscope as something you just don't start the year without. Sketchy is the resistance training program that builds that strength through three connected tools:
- Visual lessons: 1,300+ high-yield virtual lessons with recurring characters and symbols link related concepts together, so content from week two is still familiar in week twelve, not something you're re-learning from scratch
- Clinical simulations: 405+ interactive clinical simulations provide the low-stakes patient cases you need to practice differential-building and interpret labs and imaging before you're doing it live in front of an attending
- QBank questions: 9,850+ practice questions to target your weak spots in a custom test builder, with explanations that link straight back to the sketch that taught the concept
Each tool loads a little more weight onto the last. The sketch builds the memory, the case puts it under realistic strain, and the QBank question confirms it holds up when it’s exam time.
How Sketchy shows up in your preclinical years, not just the week before boards
Pharm, pathophys, micro, and biochem are the courses that hit hardest early, and they're the ones that come back on block exams and the boards. The catch is that these exams reward different things. Block exams reward depth; boards reward breadth. So the smart play early on is to learn each topic well enough to pass the block in front of you, while quietly building the broad foundation the boards will test months or years later.
Sketchy's visual lessons are built for exactly that. The recurring characters and symbols that make a concept click in week two are still doing work when it resurfaces on a cumulative exam you didn't specifically study for that week. And because you learned the material visually the first time, a 15 to 20 minute rewatch is often enough to bring back concepts from weeks or months earlier. There’s no need to re-learn from scratch.
What medical students tell us:
- 96% report higher exam scores
- 100% say Sketchy makes it easier to remember what they study
- 86% of Sketchy case users feel more prepared for clinical rotations
- 97% feel more confident building differentials
Preclinical year is also the right time to start getting accustomed to clinical reasoning itself, not just the facts underneath it. Interactive patient cases let you practice differentials and decision-making privately, with no attending or classmate watching, so you can build your reasoning ability well before you start your rotations.
How Sketchy keeps you sharp once rotations start, when study time disappears
Clinical year doesn't just shrink your study time, it changes how you have to think. In preclinical years you organized knowledge by condition. On the wards, patients arrive as chief complaints, and you have to reason from there to the diagnosis. That rewiring is hard, and it's happening while you're adjusting to the nonstop pace of rotations and exams.
Sketchy organizes its case content by rotation, so you can train exactly what each one tests. Each case starts with a chief complaint, then walks you through history, exam, differential, and management, the same sequence you'll follow on the wards.
The value these cases provide is the feedback. When your reasoning doesn't hold up, the virtual preceptor doesn't just say that you’re wrong; it shows you where the logic broke and how to work through it. You can run every case for your specialty before the rotation starts — and keep running them throughout — so the common presentations, diagnostic patterns, and clinical language start to feel familiar before your first real patient, and stay sharp by the time your shelf exam lands.
On top of that, the pharm, patho, and biochem from your first two years can fade without reinforcement, right when a patient in front of you needs that recall to hold. Because Sketchy's visual method made it stick the first time, review takes minutes between patients, not hours you don't have.
Sketchy cases target the reasoning skills shelf exams, boards, and attendings expect
Sketchy Cases focus on the skills you'll actually use on rotations and on exams:
- Diagnostic decision-making
- Imaging and lab result interpretation
- Infectious disease workup and management
- Board-style clinical vignettes
- Differential-building drills
Working through the cases for an upcoming rotation is a straightforward way to get familiar with what it'll ask of you before you start.
Get 25% off Sketchy for medical and DO students
The work you’re putting in now to get the information in your head and clinically reason from it is the strength you'll draw on when you’re on rounds, on boards, or in front of a patient who's counting on you to know the answer.
Lock in the discount before it's gone: save 25% on Sketchy with code BTS25.
Common questions
Tan, J.W.D., Oh, H.X., Krishnan, V., Neo, A.W.M., Mohamed, I.N., Mohamed Pakri Mohamed, R., Chowdhury, M.E.H., & Kumar, J. (2026). Prevalence of AI chatbot usage and factors associated with AI dependency among medical faculty undergraduate students. Frontiers in Education, 11:1750324.
Bastani, H. et al. Generative AI without guardrails can harm learning: evidence from high school mathematics. Proc. Natl. Acad. Sci. USA. 122, e2422633122 (2025).
Ke, Y., Jin, L., Ong, J.C.L., Thirunavukarasu, A.J., Car, J., Cheung, C.Y., Tham, Y.C., Ting, D.S.W., Ong, M.E.H., Compton, S., Narayan, A., Keane, P.A., Wong, T.Y., Bates, D.W., Tan, P., & Liu, N. (2026). AI-induced never-skilling in medical education. Nature Medicine, 32, 1997–2006.
Teng, D., Tan, L., Cao, Q., Xia, Y., Zhang, N., Li, J., & Zhao, D. (2026). Impact of AI misinformation on diagnostic accuracy and confidence calibration in novice medical students. npj Digital Medicine, 9:356.