My Kid Wants to Be a Doctor or Lawyer — What to Tell Them About AI

When your 12-year-old says she wants to be a doctor, the instinct is to encourage it without qualification. These are still prestigious careers. They pay well. They matter to society. All of that remains true.

But “don’t worry, those jobs are safe” is not the full truth — and kids who get incomplete information now will face harder adjustments later. The professional landscape for medicine, law, and accounting is shifting under the feet of everyone currently practicing, and it will look substantially different by the time a kid entering middle school today is working in one of those fields.

Here is what’s actually happening — and what to tell your kid.

The Incomplete Frame: “AI Will or Won’t Replace Professionals”

The debate usually gets framed as binary: either AI will take these jobs, or it won’t. Both camps have credible people arguing their side. The binary frame is the problem.

The more accurate frame is this: AI will not eliminate doctors, lawyers, or accountants, but it will eliminate large portions of what they currently do. The roles will survive. The job content is being reorganized.

A radiologist who reads routine chest X-rays to flag pneumonia will be replaced, or more precisely, will become a supervisor of AI systems that do that reading faster and more accurately. The radiologist who interprets an ambiguous finding, communicates uncertainty to a patient, integrates that finding with a patient’s personal history and values, and makes a judgment call under conditions of incomplete information — that radiologist is doing something AI cannot yet replicate.

A lawyer who drafts standard non-disclosure agreements from templates will be replaced. That’s already happening — contract drafting AI tools like Harvey, Ironclad, and Kira are in active use at law firms. A trial lawyer who reads a jury, builds trust over weeks of testimony, and makes split-second tactical decisions based on reading the courtroom will not be replaced.

An accountant who processes routine tax returns using standard deductions will be replaced. An accountant who advises a business owner on the tax implications of a complex cross-border acquisition, weighs legal risk, and recommends a structure that balances liability exposure against cash flow — that person is doing judgment work, not processing work.

This distinction matters enormously for what we tell kids.

What the Data Shows

The Goldman Sachs 2023 report The Potentially Large Effects of Artificial Intelligence on Economic Growth estimated that AI could automate tasks equivalent to 300 million full-time jobs globally. But the key finding is more granular: not jobs would disappear, but tasks within jobs would be automated.

Profession	Tasks Most Exposed to AI Automation	Tasks Least Exposed	Net Job Impact (Goldman Sachs estimate)
Radiologist	Routine image reading (pneumonia, fractures, standard screening)	Ambiguous case interpretation, patient communication, rare pathology	Role shrinks by volume, not by count
General Practitioner	Symptom triage, standard prescription refills, routine referrals	Differential diagnosis in complex cases, patient relationships, end-of-life care	Productivity increases; fewer GPs needed per patient population
Corporate Lawyer	Document review, contract drafting, legal research, discovery	Trial advocacy, client strategy, ethical judgment, negotiation	Legal services market growing; distribution shifts toward AI-using firms
Tax Accountant	Standard returns, basic bookkeeping, IRS correspondence	Complex tax strategy, M&A structuring, audit defense	Routine work collapses; advisory work grows
Auditor	Data sampling, compliance checklists, variance detection	Materiality judgment, fraud detection interpretation, board communication	Efficiency gains; smaller teams, higher complexity
Surgeon	Pre-op planning (some), post-op monitoring	Intraoperative decision-making, complex anatomy, rare cases	Augmentation, not replacement; robotic assist grows

Sources: Goldman Sachs Economics Research (2023); McKinsey Global Institute A New Future of Work (2022); Stanford HAI Artificial Intelligence Index Report (2024).

The Stanford HAI 2024 report adds context: AI is augmenting high-skill professions rather than eliminating them outright. But “augmenting” means the professionals who adopt AI tools are doing more work per hour — which compresses demand for the professionals who don’t. Fewer lawyers may be needed to produce the same legal output. The ones who survive the compression are the ones using the tools.

Medicine: What Changes and What Doesn’t

What AI is actually doing in clinical settings today

AI-powered diagnostic tools are in active clinical use, not hypothetical deployment. The FDA had approved over 700 AI/ML-enabled medical devices as of 2024, according to the FDA’s AI/ML Action Plan database. These include algorithms for detecting diabetic retinopathy from retinal scans (IDx-DR, cleared 2018), detecting breast cancer in mammograms (Transpara, widely deployed in Europe), and flagging intracranial hemorrhage on CT scans.

In radiology specifically, a 2023 study in The Lancet Digital Health found that an AI system detected screen-detected breast cancers with equivalent accuracy to two radiologists working independently — and with a 44% reduction in radiologist workload. The efficiency gain is real. The implication for the profession is also real: fewer radiologist-hours per screening study.

What AI cannot do (yet, and perhaps for a long time)

Clinical medicine at the edge of its difficulty is irreducibly human. Consider: a 67-year-old patient with atrial fibrillation, newly diagnosed kidney failure, early cognitive decline, and a family that can’t agree on her care goals. Her chart contains 20 years of notes, multiple conflicting prior diagnoses, and cultural factors her previous providers didn’t document. What’s the right management plan?

No current AI system handles this. Not because it lacks data — because the problem requires integrating incomplete information with patient values, communicating across a family in conflict, making probabilistic judgments about what this specific patient will tolerate emotionally and physically, and doing all of this while the patient watches your face for reassurance.

That is the job of a physician. AI makes it more efficient at the margins. It doesn’t replace it at the center.

The message for kids who want to be doctors

Tell them: medicine is being augmented by AI, not eliminated. The doctors who will thrive are those who understand what AI diagnostic tools are doing, can evaluate their outputs critically, and add the human layer that tools cannot provide. The ones who refuse to engage with AI tools will lose efficiency to colleagues who do. This is not a threat — it’s a description of how every medical technology has changed the profession over 100 years.

Law: Where Disruption Is Already Visible

Law is arguably the professional field where AI disruption is most advanced right now, because so much legal work is language-based — and large language models are very good at language.

Harvey (the AI legal assistant backed by OpenAI) is deployed at Allen & Overy, PwC Legal, Macfarlanes, and others. A&O Shearman reported that Harvey handled over 40,000 client matters in its first year of deployment. Kira Systems, now part of Litera, has automated contract review for years at major firms. Thomson Reuters’ CoCounsel (powered by GPT-4) is now used by thousands of legal professionals for research, contract analysis, and deposition preparation.

The McKinsey 2022 Future of Work report estimated that 23% of total work hours in legal services could be automated with current AI technology. That’s not a future projection — it’s based on technology already deployed.

What this means in practice: law firms are already doing more work with the same headcount. Junior associate positions — the traditional entry-level pipeline where young lawyers did document review and research — are under pressure. Several major law firms have reduced associate hiring while increasing use of AI tools.

But the profession isn’t shrinking. The total demand for legal services is growing, because AI makes legal services cheaper and accessible to clients who couldn’t afford them before. The composition is shifting. High-complexity work grows. Routine commodity work contracts.

A kid who wants to be a lawyer should know: learn the law deeply, learn to argue, learn to persuade and listen. And expect to use AI tools fluently — because the lawyer who can direct AI to do 10 hours of research in 20 minutes has a real competitive advantage over the one who can’t.

Accounting: The Most Direct Disruption

Of the three professions, accounting faces the most straightforward near-term disruption. Much of accounting is rule-following applied to numbers — and rule-following applied to numbers is exactly what AI does well.

TurboTax’s AI features already handle the vast majority of standard individual tax returns without professional assistance. QuickBooks, Xero, and other platforms have automated bookkeeping reconciliation. The BLS projects that bookkeeping, accounting, and auditing clerk positions — the entry-level pipeline — will decline 6% from 2022 to 2032, faster than average.

But CPAs who do tax strategy, M&A advisory, forensic accounting, and audit judgment are in a different market. The AICPA reported in 2023 that advisory services revenue at accounting firms grew 16% year-over-year — while traditional compliance services grew 4%. The profession is bifurcating, fast. The high end is growing. The low end is contracting.

The useful piece of career advice here is concrete: an accounting career built entirely on compliance and return preparation is more exposed than one built on advisory and strategy. Kids who want accounting careers should understand this and plan their specialization accordingly.

The Human Skills Argument — and Why It’s Not Just Feel-Good Advice

There’s a version of this conversation that sounds like: “Don’t worry, the soft skills will save you.” That framing undersells what’s at stake. The human skills that matter here aren’t soft.

Judgment means reasoning under uncertainty, with incomplete information, when the stakes are high. Every complex medical diagnosis, every trial, every audit of a company with a messy history requires this. It cannot currently be replicated by AI because AI systems don’t bear responsibility — and professional judgment is inseparable from accountability.

Ethical reasoning means knowing what the rules say, knowing what they’re trying to accomplish, and making a call when they conflict. A doctor facing a patient who wants a treatment that evidence doesn’t support. A lawyer whose client wants to pursue a legal but harmful strategy. An accountant who discovers that what’s technically permissible is materially misleading. These are judgment problems, not knowledge problems.

Communication — specifically, the kind that changes minds, builds trust, and works under emotional pressure — is extraordinarily hard to replicate. Cross-examination of a hostile witness. Breaking bad news to a family. Explaining a complex tax restructuring to a board that has mixed technical literacy. These require reading people and adapting in real time.

The Stanford HAI 2024 report found that human-AI collaboration consistently outperforms either humans alone or AI alone on complex professional tasks. The differentiator isn’t the AI tool — it’s the human who knows what to ask of it, how to evaluate its outputs, and what it can’t reliably do.

That’s the real message for your kid: the human skills matter more than ever, AND understanding the AI tools gives you a systematic advantage. It’s not either/or.

What Parents Should Do

Tell them the truth about what AI is changing — specifically

Vague warnings about AI disruption aren’t useful. Specific information is. Tell your future doctor: “AI can already read X-rays as well as radiologists for routine cases. Your job will be to handle the hard cases and talk to patients.” Tell your future lawyer: “AI is already doing contract review and research. Your job will be arguing, persuading, and thinking strategically.” Specific awareness beats vague anxiety.

Insist on AI literacy alongside professional preparation

This means more than “learn to use ChatGPT.” It means understanding how AI systems work well enough to evaluate their outputs critically, know where they fail, and use them strategically. A pre-med student who understands how a convolutional neural network flags anomalies in medical images is better equipped to use those tools in clinical practice than one who just learned to type prompts. For more on building this foundation, see our guide to future-proofing kids’ careers with AI skills.

Don’t discourage these professions

Medicine, law, and accounting are still excellent careers — well-compensated, socially valuable, intellectually demanding. They are not going away. The professionals who will struggle are those who resist engaging with AI tools, not those who learn to work with them. This is not unique to 2026 — every professional generation has had to adapt to new tools.

Frame the human skills as technical, not “soft”

If you call communication and judgment “soft skills,” your kid may undervalue them in a culture that prizes coding bootcamps and technical credentials. Reframe: “The ability to argue a position under cross-examination is a technical skill. The ability to diagnose a patient with five conflicting symptoms is a technical skill. These take as much development as coding.” They do.

Let them see real professionals working with AI

If you have a doctor, lawyer, or accountant in your network, ask them to show your kid how they’re actually using AI tools in their work. Real examples are more persuasive than articles. A lawyer demonstrating how they use Harvey to analyze a 500-page contract in 20 minutes, then applying their own judgment to the results, is a powerful image for a teenager considering the profession.

What to Watch Over the Next 3 Years

Medical AI FDA approvals will continue to accelerate. Watch for AI tools moving from imaging into clinical decision support — AI that helps a physician manage a complex patient, not just read a scan. This is where the most interesting human-AI collaboration questions arise.

Law firm AI adoption rates will be visible in hiring data. If major firms continue to reduce associate headcount while deploying AI, the traditional entry-level pipeline becomes more competitive. High-caliber legal education becomes more important, not less.

Accounting regulatory response to AI-generated financial work is still being sorted. The PCAOB (Public Company Accounting Oversight Board) is still developing standards for AI involvement in audit work. How those standards evolve will shape the profession’s structure.

Bar exam and medical board examination formats may shift to test AI-augmented performance rather than memorization. Several law schools are already experimenting with AI-permitted assessments. If professional licensing tests change to reflect real practice, what students need to learn changes with it.

Frequently Asked Questions

Will AI reduce the pay in these professions?

Not at the top. High-complexity work in medicine, law, and accounting commands premium pricing and will continue to do so as AI handles routine work. Compression is more likely at the entry and mid-levels, where AI tools are most substitutable. The path to the high end requires more demonstrated judgment and expertise, not less.

Should my kid choose medicine over law because medicine is “safer” from AI?

Both professions face significant AI impact at the routine end. Medicine may have a longer runway for human necessity in some subfields (surgery, psychiatry, primary care relationships). Law is seeing faster current disruption. But the long-run risk profile depends more on which subdomain your kid enters than on the profession overall.

My kid wants to be a general practitioner, not a specialist. Is that safe?

Primary care faces a nuanced picture. AI-enabled telehealth and symptom-checking tools are taking routine cases that would have gone to a GP. But complex multi-problem patients — the majority of what fills a GP’s calendar in an aging population — require exactly the kind of integrated human judgment that AI doesn’t handle. Long-term, GP demand is likely to hold; the nature of the cases will shift toward complexity.

What’s the risk of training for a profession that gets disrupted before my kid finishes their degree?

Law school takes 3 years post-grad. Medical school takes 4 years plus residency. These timelines mean a kid starting undergraduate today is 10–15 years from full professional practice. The disruption will be visible well before they finish training — giving them time to adapt their specialization. The key is building awareness now, not anxiety.

Is a law degree still worth the debt?

This depends on the school, the debt load, and the specialization. Top law schools feeding into high-complexity legal work remain financially viable. Lower-ranked schools feeding into commodity legal work — document review, routine contracts — face a more difficult value proposition as AI automates those entry-level roles. The debt calculation has always mattered; it matters more now.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Goldman Sachs Economics Research. (2023). The Potentially Large Effects of Artificial Intelligence on Economic Growth. Goldman Sachs Global Investment Research. https://www.goldmansachs.com/intelligence/pages/generative-ai-could-raise-global-gdp-by-7-percent.html
2. McKinsey Global Institute. (2022). A New Future of Work: The Race to Deploy AI and Raise Skills in Europe and Beyond. McKinsey & Company. https://www.mckinsey.com/mgi/our-research/a-new-future-of-work-the-race-to-deploy-ai-and-raise-skills-in-europe-and-beyond
3. Stanford Institute for Human-Centered AI. (2024). Artificial Intelligence Index Report 2024. Stanford HAI. https://aiindex.stanford.edu/report/
4. Salim, M., Wåhlin, E., Dembrower, K., et al. (2023). “External evaluation of 3 commercial artificial intelligence algorithms for independent assessment of screening mammograms.” JAMA Oncology, 9(10). https://doi.org/10.1001/jamaoncol.2023.3483
5. U.S. Food and Drug Administration. (2024). Artificial Intelligence and Machine Learning (AI/ML)-Enabled Medical Devices. FDA. https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-enabled-medical-devices
6. Bureau of Labor Statistics. (2024). Occupational Outlook Handbook: Bookkeeping, Accounting, and Auditing Clerks. bls.gov. https://www.bls.gov/ooh/office-and-administrative-support/bookkeeping-accounting-and-auditing-clerks.htm
7. American Institute of Certified Public Accountants. (2023). AICPA & CIMA 2023 Annual Report: Trends in Advisory Services. AICPA. https://www.aicpa-cima.com/news/article/aicpa-cima-2023-annual-report
8. Allen & Overy. (2023). Harvey AI Deployment: One Year Review. A&O Shearman press release. https://www.aoshearman.com/news/