Lawyers Are Not Risk-Averse. A Thirty-Year Dataset Says Something Worse.

Every change programme in the legal sector rests on the same assumption: lawyers are risk-averse.

Dr Larry Richard has been measuring the personality traits of lawyers for more than three decades, and on an ILTA panel on 7 August 2026 he said something that should have stopped the room. There is a measurable trait called cautiousness. It is the trait that captures how risk-averse a person actually is. When he measures lawyers on it, they produce a classic bell curve. Some are cautious. Some are middling. Some are open to trying new things.

On his instrument, lawyers score just like the general public.

His own reaction, after looking at that result year after year: why is this so consistently normal, and not elevated on the risk end of the scale, when everything about how lawyers behave says it should be?

The trait that is actually elevated

Something in the profile is skewed, but on Richard's data it is not cautiousness. It is resilience, and it runs the other way. His research puts nine out of ten lawyers low in resilience. Defensive, insecure, easily wounded, uncomfortable with criticism. It is the subject of Thin-Skinned, due from the American Bar Association in August 2026.

A low-resilience person avoids mistakes for a completely different reason than a cautious person does. The cautious person dislikes the risk. The low-resilience person dislikes what follows the mistake, which is being criticised for it.

The behaviours look identical from the outside. On his account they are not the same thing, and they do not respond to the same treatment.

The profession may have spent thirty years treating a criticism problem as a risk problem.

How do you tell caution from low resilience?

Richard offers a diagnostic that settles it, and it is disarmingly simple.

A genuinely cautious person is cautious when they are alone in a room hanging a picture on the wall. Nobody is watching. The caution is still there, because the caution is about the risk.

A low-resilience person shows risk-averse behaviour only when someone is observing them. Remove the observer and the behaviour goes away, because the fear was never about the risk. It was about the audience.

Then Richard adds the part that explains how the profession could have got this wrong. Most of us are observing lawyers when we are in the room with them. We are the observer. So we are almost always looking at the low-resilience mode, seeing it look exactly like caution, and drawing the obvious conclusion.

The obvious conclusion may be the wrong one.

What does the misdiagnosis cost a law firm?

Consider what a managing partner does when they believe their partners are risk-averse.

They build a business case. They commission a proof of concept. They find comparative data from firms of similar size. They bring the supplier in to answer objections. They present, carefully, at a partners' meeting, with evidence proportionate to the perceived caution of the audience.

Every one of those moves is aimed at reducing perceived risk. If the room's actual constraint is fear of being seen to fail, none of it lands. Worse, a partners' meeting is the highest-observation environment the firm has. It is the single worst room in the building in which to ask a low-resilience person to commit to something new.

More evidence is the correct response to caution and the wrong response to exposure.

Which would explain a sequence I hear described often enough to notice. The business case was strong. The partners agreed it was strong. Nobody said no. And nothing happened.

The shadow adoption tell

There is a second-order consequence worth testing, and it is testable inside any firm.

If risk aversion in lawyers is largely fear of observed failure, then private experimentation should be running ahead of declared adoption. Julia Montgomery, who runs change programmes on this material, described the partner who arrives on a Monday morning having spent the weekend building something with Claude and now wants it deployed at the firm. She says she has a number of them.

The prediction that follows is not hers, it is mine, and any managing partner can check it against their own firm: that partner is the visible edge of a larger group doing the same thing quietly and telling nobody.

If that holds, unsanctioned AI use is not primarily a governance failure. It is a symptom, and the symptom is legible: people will use these tools where they cannot be watched failing with them. Every policy that raises observation without raising psychological safety gives those people one more reason to hide what they are already doing.

A firm with low declared adoption and high shadow usage does not have a compliance problem. It has a safety problem wearing a compliance costume.

What actually moves AI adoption in a law firm?

Richard is specific about the mechanism, and it is slower than most firms want.

Reduce the risk of someone feeling low-resilience by authentically supporting their strengths, and not at the moment you happen to want something from them. He describes it as a long-term play: a strengths-based culture and a performance management system weighted towards bringing out the best in people rather than fixing their deficiencies. His qualifier matters. There is always room in the model for fixing deficiencies that genuinely have to be fixed. It is about the ratio.

Where the ratio sits mostly on fixing deficiencies, he says you get a low-resilience response almost anywhere you look. Where people carry a durable sense that you see them and are working to bring out their best, they feel psychologically safer, and safer people try new things.

Note what is absent from that prescription. No supplier. No pilot. No business case. On his account the thing that moves technology adoption in a law firm is a partnership culture question that has nothing to do with technology, and one that had to be answered long before anyone opened a demo.

Two firms, viewed from the outside

Take two independent firms of similar size and practice mix. Both bought the same tools in the same quarter. Both report disappointing usage.

The first firm reads its usage numbers as evidence of scepticism and responds with more training, more evidence, more mandate. Attendance is polite. The numbers do not move, because every additional session is another observed environment in which a partner might reveal they do not understand something.

The second firm reads the same numbers as evidence that nobody wants to be watched being a beginner. It runs small, handpicked, low-visibility groups. It stops publishing league tables of usage by fee earner. It changes what partners get praised for in front of their peers.

On paper the two firms made the same investment in the same quarter.

Only one of them is buying anything.

The uncomfortable part

There is a version of this that is harder to say out loud in a partnership.

If the constraint on adoption is fear of visible failure, then the person whose behaviour matters most is the one with the most to lose from being seen to struggle. In an independent firm, that is usually the senior name on the door.

Montgomery makes the structural point cleanly. When a partner stays visibly disengaged, the associates working under that partner read it as a signal about what is safe to do with that partner's work. The partner does not have to block anything. They do not have to touch a tool at all. Their silence is sufficient.

Which would mean the adoption problem in most independent firms is not distributed across the partnership at all. It is concentrated in one or two people, and it is not about their view of the technology.

For three decades the profession has explained its own conservatism with a trait that at least one long-running measurement does not support. The explanation was flattering, which is probably why it went unexamined. Caution reads as a professional virtue.

The firms that move first on this will not be the ones with the best tools or the most convincing business case. They will be the ones where a senior lawyer can be visibly bad at something for a fortnight without it costing them standing.

That is a cultural condition, not a procurement decision. And it is very hard to build quickly once you discover you need it.

Frequently asked questions

Are lawyers actually risk-averse?

Not according to Dr Larry Richard's data. On the trait that measures risk aversion, cautiousness, lawyers produce a normal bell curve and score just like the general public. The trait that is skewed is resilience: nine out of ten lawyers score low, which means they avoid mistakes because they dislike being criticised, not because they dislike the risk.

How can you tell caution from low resilience?

A genuinely cautious person is cautious when alone in a room hanging a picture. Nobody is watching and the caution is still there, because it is about the risk. A low-resilience person shows risk-averse behaviour only when someone is observing. Remove the observer and the behaviour goes away, because the fear was about the audience, not the risk.

Why does a strong business case for legal tech produce no action?

Because a business case, a proof of concept and a supplier presentation are all aimed at reducing perceived risk. If the real constraint is fear of being seen to fail, none of it lands. A partners' meeting is the highest-observation environment the firm has, which makes it the worst room in the building in which to ask a low-resilience person to commit to something new.

What does unsanctioned AI use in a law firm signal?

If risk aversion is mostly fear of observed failure, private experimentation will run ahead of declared adoption. People use tools where they cannot be watched failing with them. A firm with low declared adoption and high shadow usage does not have a compliance problem. It has a safety problem, and policies that raise observation without raising safety push that usage further underground.

What actually improves technology adoption in a law firm?

Richard's answer is a strengths-based culture and a performance management system weighted towards bringing out the best in people rather than fixing deficiencies. Safer people try new things. Note what is absent: no supplier, no pilot, no business case. Adoption is a partnership culture question that had to be answered long before anyone opened a demo.