New-model law firms are branding themselves AI First, with the thesis being: AI can learn to automate human thinking on legal issues and practices by mining legal data (such as negotiated agreements, opinions, laws, and regulations) to predict outcomes, gradually phasing the lawyer out of the picture. At Flatiron Law, however, we are not on board with that. To be clear, we believe that AI is a force multiplier that can enable talent to do more, better, faster and with less labor and less overhead.

However, for high-value, high-stakes deal work, like M&A and private equity — the domain in which we practice — you need senior talent with subject expertise to drive the AI output to ensure the requisite quality is delivered to the client. That’s why our model is Talent First, AI-powered.

The challenges of relying on AI alone are well documented: hallucinations, bias, ethics, and more. But one major issue gets overlooked — what some have called the monoculture trap, or more simply the homogenization of the practice. Because models draw from the same (or substantially the same) source data, AI agents are left to analyze clauses or negotiate terms without human wisdom. That means judgment and direction will always converge on whatever consensus outcome the data produces. Everybody gets the same homogenized outcome — or what can be called breeding mediocrity.

Commonality of outcome may be just fine for commodity work at the low end of the legal spectrum, such as repeatable commercial contracts like MSAs, SaaS Terms, and NDAs. Clearly, you do not need to break the bank to marginally improve your standard NDA; but this does not work for high-value, high-risk transactions, regulatory analysis, litigation, or any situation in which talent, judgment, and expertise are required.

Players in high-stakes dealmaking are not going to settle for the lowest commonly denominated, homogenized outcome that everyone else gets. Clients want creative, superior outcomes that maximize the value of their negotiating position — whether that’s strong or weak. And they’re willing to pay for results. Indeed, firms like Wachtell Lipton are paid for value, not by the hour, and clients gladly pay it.

Those kinds of outcomes require senior talent. And while AI can help by accelerating decision-making by identifying optionality and analysis. At the end of the day, however, talent drives the creative process.

The 7 risks of AI First in high-stakes deal making

The risks associated with an AI First approach extend beyond accuracy concerns, often raising broader questions about judgment, differentiation, training, quality control, and professional responsibility. Here are seven risks for which to be watchful:‍

1. Race to mediocrity

In an adversarial legal system, the value of advice is partly a function of how it differs from what opposing counsel produces. When both sides converge on the same canonical analysis, what looks like consensus is the loss of one side’s edge. For deal work, the practical translation is stark: The difference between getting market terms and getting the deal to which your leverage entitles you. Worse, the lawyers with the better-but-less-common argument may now find that their reasoning sounds unusual or non-market against the LLM-generated baseline — and be tempted to abandon it.

2. The "rational" answer that breaks the deal

The subtler danger is not that the model is wrong, but that it is right in the abstract and wrong for the deal. Many LLM models optimize for the rational position — the one the data says is the best outcome for the client — and will advance it regardless of where the leverage actually sits. However, one of the most important advantages a deal lawyer has is the ability to calibrate reality for the client: What is realistically achievable? What is not? And where to spend finite negotiation capital given the parties' leverage?

Models do not do this. Absent careful prompting by a subject-matter expert, they do not weigh leverage, goodwill, or timing risk — and left unfiltered, their rational output can do more damage than good, stressing deals and burning goodwill that a seasoned lawyer would have protected.

We have seen this failure mode more than once. Clients have used AI to generate comments late in a negotiation — comments that, in an ideal world, would have advocated for better terms, but in the real world would have stressed (and in one case nearly broken) the deal. This is where Talent First earns its keep, by triaging model output and reconciling it with reality. The intervention of subject-matter experts kept a rational-on-paper position from producing an irrational outcome.

3. The bespoke becomes generic

Sophisticated transactional practice depends on negotiated, idiosyncratic provisions, such as unusual carve-outs, novel indemnities, calibrated reps, warranties, and creative earn-out structures. Those are exactly the provisions least likely to be generated by an LLM trained on model contracts. And over time, as LLM-drafted contracts become the next generation’s training data, the model contracts narrow further simply because there is no equilibrium other than convergence.

The pessimistic case is not that bespoke drafting disappears; rather, it’s that bespoke itself gets outsourced to the model, which produces generic versions of bespoke, without any of the craft or original thinking that made bespoke worth paying for in the first place. Thus, the overall value of the firm offering is degraded.

4. The fluency trap: Abdicating quality control

The fluency of the LLM models can be a trap itself, offering output that reads well and inviting the assumption that it is well-reasoned. This can tempt firms to relax the layer of senior review that they historically deployed to ensure quality control. That is a dangerous concession. An output that reads so well may be flawed for all of the reason discussed, and its polish is exactly what lets those flaws slip past review.

5. Junior lawyer de-skilling

Junior lawyers historically built their legal judgment by doing the cognitive work that LLMs now perform, such as reading every cited case, identifying the doctrinal architecture, drafting the first analytical pass. What AI removes from junior lawyers is not the typing but the thinking — the formative practice that develops legal judgment. Today’s senior lawyers can use LLMs effectively because they were trained without them, and they can recognize when an output is wrong or shallow. The lawyers being trained today don’t have that luxury because the models are now doing that work in their place.

6. Cascading failures

When both sides rely on the same model, no one is positioned to catch the other's errors — and errors don't tend to aggregate, they cluster. The reality is that the models make errors, and they always will. The research is blunt about it: Stanford's Hallucination-Free? study found that even the leading legal-research tools hallucinated on a material share of queries — roughly one-third for Westlaw's AI, and 1-in-6 for Lexis-Plus. That’s lower than general-purpose models, but nowhere near zero.

Sanctions are the consequence, and they keep escalating — from monetary penalties on law firms exceeding $100,000, to outright dismissal of cases with prejudice. Not to mention the ruling from the U.S. Sixth Circuit Court of Appeals that no lawyer may cite anything that they have not personally read and verified. This is not anecdotal; indeed, one tracker now catalogues more than 1,300 tainted filings worldwide.

7. Ethics and governance exposure

The American Bar Association’s Formal Opinion 512, released in July 2024, applies six existing duties to generative AI use: competence, confidentiality, communication, candor, supervision, and reasonable fees. Yet competence is the one that our monoculture most directly implicates: If competent representation comes to require LLM use, lawyers become structurally exposed to those models' failure modes. Confidentiality compounds the problem, since shared self-learning tools mean shared failure modes.

Since the release of that opinion, the ground has shifted from guidance toward enforcement. More than two dozen state bars have issued their own AI guidance, and the California Supreme Court has directed the state bar to move AI duties into its enforceable Rules of Professional Conduct. Courts increasingly require affirmative disclosure through standing orders, and bar authorities have begun disciplining lawyers for unverified filings. The direction is unmistakable: The duty to verify is becoming non-delegable and independently enforceable and saying, “The model produced it!" is no defense.

The non-homogenized approach

Our recommendation is not to retreat from LLM use, but to deliberately manage the model output with experts, what we’ve called Talent First. This keeps experienced human judgment — including the contrarian instinct to challenge a clean-sounding answer — at the center of the workflow, and it holds senior talent responsible for quality control rather than delegating that role to the model.

That means measuring the model's output against the realities of the deal by determining what the leverage actually supports, what advances the client's position, and what merely reads as rational. It also means refusing to let cost pressure flatten the firm's analysis to the model's median.

We recommend four disciplines to follow:

Build devil’s-advocate discipline into the workflow

Lawyers should not ask LLMs for the best argument, they should ask for the strongest argument against the obvious answer. Disciplined adversarial prompting alone will not solve generative monoculture — nothing at the prompt layer will — but it identifies more diverse output than default prompting, and it trains lawyers to keep looking for the contrarian read.

Capture your distinctive analytical patterns

Your firm’s deal teams undoubtedly have characteristic approaches, such as particular ways of structuring indemnities, distinctive theories of damages, non-obvious doctrinal frames, and signature negotiating moves. Those are precisely what monoculture erodes. To avoid that, capture and operationalize these characteristics so that training lawyers and your systems will be done on your edge — not the median of the model.

Tier work by monoculture sensitivity

Not every matter is created equal. High-volume commodity work (form NDAs, routine compliance, standard sub-doc review) tolerates convergence and benefits from speed. But high value, high stakes deal work (like M&A) requires LLM use with mandatory human review by subject matter experts. At the top of the legal work product food chain, LLM output should be a starting hypothesis to be triaged against the realities of the deal, argued against, and not relied on.

Preserve the cognitive load that builds junior judgment

Junior lawyers are the future of your firm. Any firm that stops developing them is quietly deciding not to have a next generation, and it will slowly evaporate. While juniors may no longer be the profit centers they once were, they remain critical assets to train and prepare even in the age of AI. Their development is not overhead; rather, it’s an investment in the firm's future. That is why firms should resist the temptation to replace junior research and drafting entirely with LLM output. Some training tasks remain genuinely cognitive, such as first-pass analysis without AI, followed by AI comparison, followed by senior review. The efficiency cost is the training investment.

One strategy that we have found useful to remedy this training gap is to create new virtual environments that replicate the real-live deals upon which juniors historically trained. Simulators can give juniors repeated, consequential reps at the judgment call itself, allowing them to interact with virtual clients, oppose counsel and other stakeholders before the stakes are real.

This practice also restores the one thing an AI-assisted workflows tend to strip out: the chance to exercise judgment under realistic pressure. In fact, at Flatiron Law, we have built, together with other collaborators, our own AI agentic negotiation-training platform, which we focused first on M&A.

For a fuller treatment of how firms can escape the monoculture trap—and a path forward—read the complete article on Flatiron's website