By Stuart Kerr, Technology Correspondent, LiveAIWire

Professional identity is already measurably harder to defend. Fifty percent of lawyers now call AI a major threat to the unauthorised practice of law, up from 36 percent just a year earlier, according to the Thomson Reuters Institute’s 2026 AI in Professional Services Report, which surveyed more than 1,500 professionals across 27 countries. The same survey found that, compared with 2025, a higher share of professionals now believe AI will have a major impact on jobs, billing, revenue, and even the long-term need for legal, tax, and accounting professionals as a whole.

That anxiety is not really about AI replacing professionals outright. It is about AI replacing the tasks through which professionals have always developed their expertise in the first place, and that shift changes what it means to be a professional at all, along with what a credential is actually certifying once the work behind it has changed. Law, medicine, accounting, architecture, and engineering built their authority on accumulated experience. Now the experience itself is going missing.

The Ladder Professionals Climbed On Is Being Removed From Underneath Them

Professional development has always depended on a specific structure. Entry level practitioners perform routine, codified tasks under supervision, gradually building the pattern recognition and judgement that separate expert performance from novice guesswork. A junior lawyer drafts standard documents, researches precedent, and reviews contracts under a partner’s eye, developing through repetition the instincts that later become independent professional judgement. A junior accountant reconciles accounts and prepares returns, learning through volume the detail orientation that underlies real accounting expertise.

AI is efficient at exactly these routine, high volume tasks. Contract review, legal research, document drafting, financial data processing, diagnostic screening, and architectural drafting are all capabilities where AI tools already perform at levels that compress or eliminate the junior workload professionals were once trained through. Junior hiring is already falling in AI-exposed professions, as LiveAIWire’s analysis of what the 2026 employment data actually shows about AI and job displacement found, with entry-level positions in law and finance among the hardest hit.

But the deeper consequence is structural rather than immediate. The pipeline through which expertise develops is being disrupted at its foundation, which risks producing a generation of senior professionals with less tacit expertise than their predecessors, in exactly the domains where tacit expertise is what clients pay for.

Why This Is an Identity Threat, Not Simply a Career Risk, and What It Means for You

Research by Ekaterina Jussupow, Kai Spohrer, and Armin Heinzl, published in JMIR Formative Research, found that AI can threaten professional identity along two distinct dimensions: threats to professional capabilities and threats to professional recognition. Their survey of medical students and physicians found that students experienced significantly stronger identity threat and resistance to AI than the experienced professionals did, suggesting that those who have not yet completed the expertise pipeline feel the threat most acutely, while those with deep expertise already banked feel it less.

What this means for you depends on where you sit in that pipeline. If you are early in a professional career, the threat is compounding: AI can already do tasks you expected to spend years learning, and the social status attached to expertise is declining just as you are meant to be building yours. If you are established, the research suggests you are somewhat insulated, but only if your expertise is genuinely deep rather than assumed. Recruiting the next generation now requires professions to answer an uncomfortable question: what is the value of years spent developing expertise in a domain where AI produces competent output in seconds?

How the Professions Are Rewriting Their Own Rules in Real Time

The professions are responding with varying urgency. The AICPA’s Profession Ready Initiative, launched in February 2026, is the most systematic response so far: a research-backed effort to redefine the skills early-career CPAs need, built on the premise that the old skill development pathway no longer works on its own. The AICPA is framing 2026 as the year the T-shaped professional, combining deep domain expertise with broad AI fluency, becomes the standard rather than the exception.

In law, the response is less coordinated but no less urgent. The National Law Review’s survey of 85 legal professionals, published at the start of 2026, found that the firms most likely to thrive are those that have already built defensible, metrics-driven workflows grounded in real governance, treating AI adoption as a workflow engineering problem rather than a headcount reduction opportunity. The International Bar Association has identified training as a key priority, both in teaching staff to use AI and in coaching associates so they still develop, in the IBA’s words, well-rooted expertise by the time they reach senior roles.

That second priority, making sure AI adoption does not hollow out the expertise pipeline, is the most critical and the most neglected part of the response.

What Professional Schools Keep Getting Wrong About Their Own Purpose

Professional schools are facing an uncomfortable question. If AI can perform the tasks that professional education teaches, what is professional education actually for? The honest answer is that it was never primarily about teaching specific tasks. It was about developing the analytical frameworks, ethical reasoning, and professional judgement that let practitioners apply expertise to situations no training could fully anticipate. The tasks were always the vehicle, not the point.

That means the tasks used to build professional capability need to change, not that the capability itself has become obsolete. Legal education that requires students to evaluate AI-generated analysis and find its errors builds the evaluative skill that matters most in AI-augmented practice. Medical education that requires students to challenge AI diagnostic suggestions with their own clinical reasoning builds the same critical instinct. LiveAIWire’s reporting on why students are quietly replacing Google search with generative AI found a parallel risk playing out earlier in the pipeline, in undergraduate research habits, which suggests professional schools are inheriting a cohort that has already outsourced some of the cognitive habits they now need to rebuild.

The Clients Who Have Already Decided What Your Expertise Is Worth

Part of this challenge is already being resolved by clients whose behaviour is changing faster than professional institutions are. Corporate legal departments that can use AI for routine research and contract review are cutting external legal spend on that work and expecting outside counsel to move faster at lower cost, rather than paying human-expertise rates for work AI now does. Accounting clients using AI tax preparation for standard returns are retaining human accountants specifically for complex situations, planning, and judgement calls. The market is already sorting professional work that adds value beyond what AI can produce from work that does not.

The capability commanding a premium now is not the capability to perform tasks AI can already perform. It is the capacity to exercise judgement, hold client relationships together, navigate ethical complexity, and take personal accountability for outcomes. LiveAIWire’s coverage of why independent workers are both the most exposed and the most adaptable group in the AI labour market found the same pattern playing out among freelancers, where the professionals who professionalised their AI use, rather than simply speeding up commodity output, are the ones commanding higher rates.

Why the Junior Hiring Collapse Will Cost More Than It Saves

Stanford Digital Economy Lab research published in November 2025 found that, since the widespread adoption of generative AI, employment for 22 to 25 year olds in the most AI-exposed occupations, a category that includes entry-level accounting, has fallen by roughly 16 percent relative to less exposed roles and more experienced colleagues in the same jobs, even after accounting for firm-level shocks. That is not simply a story about fewer jobs today. Junior professionals are not only performing work that generates current revenue. They are developing the expertise that makes them effective senior professionals a decade later, and building the institutional knowledge that keeps organisations functioning over time.

When the junior pipeline contracts sharply, the consequences are not felt immediately. They surface a decade later, when firms that cut junior hiring in the 2020s discover a shortage of experienced mid-career professionals and a gap in the knowledge transfer that junior-to-senior progression used to provide. The IBA’s recommendation that firms invest in coaching associates, specifically so AI adoption does not hollow out the pipeline, is a recognition that short-term cost savings from reducing junior work do not compensate for the long-term capability loss if the development system breaks at its foundation.

The firms that have understood this are growing junior cohorts while using AI to make those juniors more productive, rather than using AI to replace the juniors the firm’s long-term health actually depends on.

The Career-Long Squeeze Nobody Warned Mid-Career Professionals About

This disruption is not only a junior hiring problem. It is also a mid-career problem, as professionals who built their expertise in pre-AI workflows are now required to adapt to AI-augmented ones that change the nature of the work itself. New entrants can be trained for an AI-augmented workflow from day one. Mid-career professionals have to transfer capabilities built in one context into a fundamentally changed one, which requires both the technical work of learning new tools and the harder psychological work of reconceiving a professional identity in a domain where the tasks that defined it are being automated.

The identity threat research found this reconception is hardest for professionals who draw their sense of professional worth primarily from technical expertise rather than from the relational, ethical, and judgement-based dimensions of the work. LiveAIWire’s reporting on who thrives and who is left behind as AI reshapes the workplace and on why Gen X is being left out of workplace AI training both point to the same structural pattern: the professionals struggling most are often not the least skilled, but the ones given the least structured support in redirecting skills they already have.

The Professional Identity That Actually Wins From Here

The most constructive frame here is not resistance to AI, and it is not uncritical adoption either. It is a deliberate reconstruction of professional identity around the capabilities AI amplifies rather than replaces. The T-shaped professional the AICPA describes, combining deep domain expertise with broad AI fluency and strategic capability, is not a compromise between old professional identity and new AI capability. It is a genuinely stronger identity for an AI-augmented world, one that can do more than either a human expert without AI tools or an AI system without human domain expertise.

The professionals building that identity now, investing in deep expertise while developing genuine AI fluency rather than surface tool use, evaluating AI outputs critically, catching their failures, and directing AI capability toward problems that actually matter, are the ones who will command the premium professional expertise has always earned. That premium simply requires more than it used to, and the professions still figuring that out in real time are the ones worth watching closely over the next few years.

About the Author

Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, emerging technology, and their impact on business, society, and everyday life. LiveAIWire publishes original AI journalism every weekday at liveaiwire.com.