There’s a low-grade panic circulating in conversations with legal teams: if AI handles first drafts and research arrives pre-digested,  how are junior lawyers supposed to learn? 

A recent survey by LexisNexis found that 72 percent of respondents believe junior lawyers may struggle to develop deep legal reasoning and argumentation if AI supersedes traditional research and drafting tasks. 

The concern is understandable. For decades, the model was simple: you learned by doing and getting it wrong.  Over time, you knew which clauses could never be missing from an SPA, and you developed an intuitive sense of what was “market.” 

But if repetition and visible mistakes reduce, what happens to learning?  

Perhaps we’re focusing on the wrong loss. Automation doesn’t remove skill development. It relocates it. 

Repetition was never the goal. It was just the delivery mechanism for learning. And if that mechanism changes, the learning does not vanish. It shifts. 

Expertise develops when lawyers are required to interrogate outputs, spot risk, make trade-offs, understand the consequences. Those capabilities don’t disappear. They simply become the work. 

 Automation changes the mechanics. It does not remove the need to think. 

From mechanical repetition to structured thinking  

When drafting platforms automate structure and logic, something important happens. The mechanics become visible.  

Answer one question differently and the contract reshapes itself on screen. The logic becomes visible and the impact of each answer is immediate. 

That accelerates a different kind of learning. Junior lawyers are not just producing text. They are observing structure, comparing positions, and seeing the consequences of decisions. 

In a glance, they can see: 

• Why is this clause structured this way? 

• Why does it change if I alter that assumption? 

The shift is subtle but significant. The foundation moves from mechanical execution to structured thinking. 

There is an uncomfortable truth underneath this. Volume was never a reliable proxy for expertise. Many lawyers drafted hundreds of agreements without ever stepping back to see portfolio-wide patterns or recurring negotiation pressure points. 

Automation reduces repetition. It also makes patterns explicit. 

Contracts as systems, not documents 

As drafting becomes more structured, contracts start to look less like static text and more like mapped positions within a system. 

Negotiation trends become visible across matters. Risk concentrations can be observed across portfolios. Accepted positions and fallback language can be analysed rather than remembered. 

The interesting questions shift from wording in isolation to design at scale. Not “is this clause perfect”, but “what happens when we take this position repeatedly?” 

Exposure to these patterns earlier in a career can deepen understanding faster than years of isolated drafting ever did.  

The real risk is passive use 

Automation is not the threat. Passive use is.  

If juniors treat AI outputs as answers, they’ll plateau fast. If they use automation to skip thinking, the skill gap will widen sharply. 

But if AI is used as a testing ground, comparing clauses, exploring alternative structures, stress-testing risk positions, their development accelerates. 

The key distinction is not between manual drafting and AI drafting. It is between passive consumption and active interrogation.   

AI can be used as a thinking partner, not a shortcut. The firms that benefit will be the ones that design training to make that the default. 

What this means for training 

If expertise no longer develops automatically through volume, firms need to be intentional about how it develops instead. That does not require abandoning tradition. It requires redesigning it. 

Three principles matter. 

First, make reasoning visible. 
Training should focus on why positions are taken, not just whether they are accepted. Structured review conversations, explicit articulation of trade-offs, and exposure to decision logic matter more than redline volume. 

Second, expose patterns early. 
Junior lawyers should see how risk and negotiation positions play out across portfolios, not just within a single agreement. Pattern recognition accelerates when context widens. 

Third, reward interrogation over speed. 
If AI is part of the workflow, juniors must be encouraged to question outputs, compare alternatives, and stress-test assumptions. Expertise compounds when curiosity is reinforced, not bypassed. 

Automation changes the mechanics of drafting. It does not diminish the importance of developing expertise. It simply demands that firms design for it more deliberately.