Generative AI has arrived inside BigLaw carrying a question that no one is quite ready to answer out loud: if the technology can do what junior associates do, what exactly are junior associates for?
It is the most uncomfortable conversation in the legal profession right now and it is only getting louder from what we hear.
For decades, major law firms operated on a clear and largely unquestioned model. Junior associates worked punishing hours on research memos, due diligence reviews, contract markups and first drafts. The work was repetitive, often unglamorous, and absolutely central to two things: generating billable revenue and building the technical foundation of a legal career.
Generative AI is now compressing that model in ways the profession is only beginning to reckon with.
The Economics Have Already Shifted
Firms are moving fast. Platforms built on systems from Harvey AI, OpenAI, Thomson Reuters and LexisNexis are being integrated across practices ranging from M&A to litigation support. The pitch to clients is straightforward: faster output, lower cost, greater efficiency.
The pitch to associates is more complicated however.
A fifth-year associate equipped with AI-assisted drafting tools can, in some practice contexts, produce work that previously required two or three juniors working across several days, which is already occurring across transactional and research-heavy practices.
When leverage changes, staffing models are going to follow. That is the part making the associate class uneasy, and understandably so.
The Training Problem No One Has Solved
Here is the tension that sits at the centre of this debate: the repetitive work that junior lawyers find least satisfying is often the same work that builds real legal ability.
For instance, research memos teach issue spotting and due diligence teaches pattern recognition. First draft submissions also teach how arguments are structured, where they break down, and what a partner actually means when they say “tighten this up.”
If all this is handed to AI before a junior lawyer has engaged with the underlying substance then law firms risks producing lawyers who can supervise technology before they have mastered the work itself.
Several prominent legal education commentators and professional responsibility scholars have flagged this directly.
The American Bar Association places competence obligations squarely on the supervising lawyer — not on the tool. If the supervising lawyer’s foundational skills were themselves built on AI output rather than deep engagement, competence becomes a cascading problem.
What Firms Are Doing
Law firm managing partners publicly maintain that AI will “augment rather than replace” lawyers — and for now, that framing holds. No credible analysis suggests mass associate redundancy in the near term.
But the profession is already seeing quiet structural changes:
- Reduced headcount growth at some firms, particularly at the junior end
- Increased scrutiny of associate utilisation and productivity metrics
- Early experimentation with value-based pricing and fixed-fee AI-assisted work
- A growing expectation that even first and second-year associates will demonstrate AI literacy
The billable hour is not going away. But its relationship to headcount is becoming less linear than it has been for the past fifty years.
What This Means for Associates Right Now
The lawyers most likely to thrive in this environment are not those who resist AI adoption, nor those who use it as a substitute for legal judgment. They are the ones who develop fluency with these tools while maintaining the analytical rigour that AI cannot replicate.
That means using AI to accelerate research while still interrogating its outputs. It means understanding the limits of large language models in legal reasoning — including their tendency toward confident-sounding errors. And it means building the client-facing, judgment-heavy skills that partners have always valued, and that no platform currently on the market is close to replacing.
The old promise — work hard enough and progression follows — has always been partly fiction. But AI has made the uncertainty more visible. The profession is not in crisis. But it is changing faster than most firms are willing to admit, and faster than most associate training programmes are designed to accommodate.
The question of who teaches the next generation of lawyers — and how — is one the profession cannot outsource to a language model.
LawFuel covers legal industry news and analysis for lawyers and legal professionals. Follow us for updates on legal technology, career strategy and law firm trends.
