The 3 profiles of lawyers in the face of AI: Are you the one who iterates or the one who waits for the ruling?

By Elena Irazabal

The arrival of Gemini 3 has encouraged me to write about something I've been observing for some time: the different ways in which legal professionals are relating to artificial intelligence.

It's not a particularly popular read, because it focuses more on our attitudes than on the promises of technology, but I think it's necessary.

Beyond the technology itself, what interests me is the attitude with which we approach it. Because we don't all start from the same place, nor do we react the same way to its limitations or possibilities. Over these past months, I've identified three very distinct profiles.

1. The professional who learns by iteration: trial, error, and judgment This group has, by far, developed the healthiest relationship with technology. They don't expect miracles or perfect results the first time, because they understand that AI operates within margins and limitations that must be explored. They test models with real cases, repeat consultations with slight variations, and compare the results with their own legal knowledge. From there, they learn to distinguish when it's worthwhile to use AI and when it's preferable to approach the matter in a traditional way. Their greatest strength is that they don't demand that the machine work for them, but rather with them: they conceive of the interaction as a collaborative process. They maintain critical distance, review the output, and don't seek an infallible oracle, but rather capable support that they know they must supervise with care and professional judgment.

2. The professional who wants to delegate everything… and gets frustrated A profile born from an unrealistic expectation: they imagine an AI capable of drafting, analyzing risks, and grasping nuances like a seasoned partner with a single click. Their expectation is built on the promise of total automation, as if the tool could absorb legal reasoning without human guidance. When reality clashes with this fantasy, and the system falters or is inaccurate, they oscillate between anger and disappointment, blaming the tool as if it had failed to fulfill an obligation. Their mistake is seeking a replacement instead of an extension. They expect miracles and receive imperfect drafts that require intervention. They fall into the paradox of automation: they end up working twice as hard to correct errors that, due to a lack of active supervision, they didn't see coming or didn't want to anticipate. What should have freed up their time ends up creating more of a burden because they relinquished control.

3. The professional who needs AI to fail to feel secure A classic roundtable scenario. They scrutinize the tool, waiting for it to fail, not to correct it, but to celebrate it like a trophy proving that "they'll never replace us." They relate to the system with suspicion and a desire for confirmation, searching for that error that will reaffirm their position. For them, a hallucination is not a technical error or an adjustable limit, but ontological evidence that the machine is inferior; a psychological defense mechanism that allows them to preserve their professional identity. While the first group asks, "How can this help me?", this one remains trapped in "How do I prove I'm better?" Their interaction with AI thus becomes an exercise in self-validation, not exploration, and they miss the opportunity to discover how technology can complement their work without calling its value into question.

So where to begin?

To position yourself in the first group—the one that iterates and builds upon its own criteria—and avoid the frustration of those who expect miracles, you have to stop putting the cart before the horse. I recently analyzed why obsessing over the latest AI is a trap if you don't understand the fundamentals. If you want to stop chasing novelties and build a solid foundation that will serve you today with Gemini and tomorrow with whatever comes next, the roadmap is here:

[WHAT YOU SHOULD (REALLY) LEARN ABOUT AI IF YOU ARE A LAWYER]

Because, in the end, the difference between a lawyer who “uses AI” and one who understands it begins with knowing what's under the hood.