The Question Is No Longer Whether to Use AI. It Is Which Tools Deserve Your Trust.
A recent editorial article in Nature Medicine asks for evidence of the value of medical AI. Evidence of clinical outcomes has been the dominant paradigm for adoption of technology within hospital systems and it has it's value. The problem is that thinking of AI as a constellation of single tools as we would with traditional medtech, misses the truly revolutionary power of AI. In this article, Dr. Sina Bari explores AI using the context of technology stack and argues that the approach is more analogous to turning from a community hospital to an academic powerhouse then it is like adopting a new tool.
Are Doctors using AI?
According to the Doximity 2026 State of AI in Medicine Report, 63% of U.S. physicians now actively use AI, up from 47% just a year earlier. Among those users, 74% report daily use. Adoption has moved past the early-adopter phase into something closer to infrastructure.
But adoption is not integration. Most physicians I talk to are using one or two tools in isolation. An ambient scribe here, a literature search tool there. What they lack is a stack: a deliberately assembled set of AI tools that covers the three domains where physicians spend the most non-clinical time. Research, writing, and practice management.
What a Personal AI Stack Actually Looks Like
A personal AI stack for doctors is a curated set of AI tools, selected by the individual physician, covering three functional layers: knowledge retrieval, content generation, and operational automation. The key distinction from institutional AI deployments is control. You choose the tools. You set the boundaries. You decide where AI output requires your review before it reaches a patient, an editor, or an insurer.
This matters because, as the AMA's augmented intelligence framework makes clear, governance should be risk-proportional. Drafting a case report is not the same risk as generating a differential diagnosis. Your stack should reflect that gradient.
Layer One: Research and Evidence Retrieval
Literature search is the most common AI use case among physicians at 35% of all AI-assisted tasks. I use large language models with retrieval-augmented generation for rapid evidence synthesis, querying with precise clinical parameters rather than broad questions. Narrow queries anchored to PubMed-indexed literature produce verifiable results. Broad queries produce confident-sounding guesses.
The verification step is non-negotiable. Every claim gets traced back to the original abstract. If a citation does not resolve to a real paper, I discard the entire output. Hallucination is not a partial phenomenon.
Layer Two: Medical Writing
Physicians use AI for writing across clinical documentation, patient education, prior authorization letters, and manuscript drafts. My approach: AI handles structural scaffolding and formatting. I supply the clinical substance. AI writes the first 60%. I write the last 40%. That last 40% is where clinical nuance lives.
Ambient scribes now represent 29% of physician AI use cases. In my experience, quality varies enormously by specialty and encounter complexity. A straightforward follow-up transcribes well. A complex surgical consultation with multiple decision points still demands significant editing.
Layer Three: Practice Management
Administrative burden is measurable: the healthcare system has seen roughly 3,000% growth in administrative positions over recent decades versus approximately 150% in physician supply. AI for scheduling, billing, and prior authorizations addresses this asymmetry directly.
I evaluate these tools on three criteria. Does it integrate with existing systems without creating a parallel workflow? Does it maintain HIPAA compliance with auditable data handling? Does it actually reduce time spent, measured in minutes per week, not in marketing language? Most tools fail that third test. They shift work rather than eliminate it.
Assembling Your Stack
Start with research and writing tools. These have the most forgiving error tolerance. Limit your stack to five or fewer tools; every addition introduces a new interface, login, and data policy to evaluate. Audit quarterly, because the model behind your favorite tool in January may be deprecated by June.
As a Stanford-trained surgeon and physician-founder, I have learned that the physicians who benefit most from AI are not those using the most tools. They are the ones who have clearly defined what they will and will not delegate to a model. Clinical taste, the ability to distinguish adequate AI output from excellent clinical communication, is the differentiating skill.
The 2025 Offcall Physicians AI Report found that 84% of physicians say AI improves their performance, but over 80% are dissatisfied with organizational deployment. That gap is exactly where the personal stack becomes essential: a way to reclaim agency over your own workflow while the institutions figure it out.
Frequently Asked Questions
What AI tools are doctors actually using in 2026?
The most common tools fall into three categories: ambient scribes for documentation (29% of use cases), literature search tools like OpenEvidence (35%), and administrative automation for billing and prior authorizations. The Doximity 2026 report shows 74% of physician AI users engage daily, with family medicine leading at 88% daily use among adopters.
How do physicians use AI for medical writing without compromising accuracy?
Effective physician AI writing follows a split workflow: AI generates structure and first-draft prose while the physician supplies clinical substance and performs final verification. Every clinical claim, dosage, and recommendation is checked against primary sources before publication. The discipline is treating all AI output as a draft, never a finished product.
What is Dr. Bari's approach to evaluating AI tools for clinical use?
Dr. Bari applies three criteria: seamless EHR integration without parallel workflows, verifiable HIPAA compliance with auditable data handling, and measurable time savings confirmed by staff feedback after a two-week trial. Tools that shift work rather than eliminate it do not survive the evaluation regardless of marketing claims.
Why are small practices adopting AI faster than large hospital systems?
Independent physicians retain decision-making authority over their tooling and can trial, evaluate, and discard tools in weeks. In large systems, 47% of physicians report institutional AI policies are confusing or still evolving, creating months of committee-driven friction that delays deployment and quality control alike.
What is the biggest risk of building a personal AI stack for medical practice?
Uncritical reliance on AI outputs, particularly in evidence retrieval where models still fabricate citations. The Doximity 2026 data shows 71% of physicians cite accuracy and reliability as their top concern across all specialties. Mitigation requires treating every AI output as a draft requiring human verification before clinical application.