News
By now, every company lawyer has been told that using AI calls for critical thinking. But critical thinking about what, exactly? About AI’s output, obviously. About the tools and one’s own use of AI, fair enough. But about the company’s AI choices? About what AI is doing to society? At this point, heads start to shake. The last item in the list might not be the lawyers’ job (or is it?). The interesting question is precisely where you draw the line. Deciding what falls outside the scope of your thoughtful consideration is a significant act of critical thinking!
Checking the output
The first and most obvious object of critical review is AI output.
Hallucinations are a familiar embarrassment; even elite law firms keep walking into them. The subtler trap is ‘misgrounding’: the AI cites a real source while quietly misrepresenting what it says. This one is harder to catch because everything looks right.
The necessary verification does not stop at substance. Form matters too, particularly if one wants to avoid the AI stylistic hallmarks that tend to flatten a text and make it appear AI-generated, which can undermine the credibility of both the text and its supposed author. For a lawyer, whose authority and standing as a business partner rely on trust, this is not a cosmetic point.
The tools and their integration
A fair amount of magical thinking still surrounds legal AI tools.
Some imagine one miracle platform that will run the entire department: litigation, contract review, templates, risk reporting, managing outside counsel, you name it. Others underestimate the work required to integrate the tools (use cases, playbooks, training…) or misjudge their organisation’s ability or willingness to change. Because of budgets, competing priorities, and IT policies, there might be a wide gap between what the demo showed and the actual setup.
Your own use of AI
Critical thinking applies just as much to how you (do not) use AI yourself. Lean too much in one direction and your cognitive abilities might atrophy, like a muscle you have stopped using. Lean too much in the other direction, and you’ll miss amazing opportunities to transform how you work, improve your output and learn new skills.
A small example. You need to write an email in a language you only half-speak. You can write in your own language, have AI translate it, and send the result untouched, or try your best in the target language, ask AI to suggest improvements, and then make the changes yourself. The second option costs you ten extra minutes, but it sharpens your skills every time you do it.
This area of critical thinking is a matter of personal responsibility. No one is making these trade-offs for you. It is up to you to balance short-term efficiency and long-term skill-building.
The company’s choices
Beyond your own use, you could have an opinion about the choices your company is making about AI (or failing to make). You are, of course, expected to do so from a legal and regulatory compliance perspective, but nothing stops you from broadening your view beyond mere legal considerations to more strategic ones. For example, you could suggest more caution when the company is rushing blindly or ring the alarm when doing nothing has become your company’s AI strategy.
Society, and where you draw the line
Critical thinking can also extend to the broader societal consequences of AI: dependence on a US-dominated oligopoly, environmental damage, cybersecurity exposure, shrinking job opportunities for young people, rising electricity costs, and so on.
Here, though, the duty of company lawyers is less clear. One could argue that these questions belong to governments and society at large, not to company lawyers, as they have neither the mandate nor the leverage to act on them.
Critical thinking on these matters is often switched off in companies. “Yes, it’s true, there are plenty of serious and important problems with AI”, a young lawyer told me, “but in practice, we have no choice. We just go for it, full speed.” In some organisations, the subject is not only set aside, but even taboo. Those who bring it up are dismissed as alarmists who don’t grasp what is really at stake, as if the underlying risks did not exist, even though they are raised by the very people who built these systems.
Whether you, as a company lawyer, will engage with any of this is a genuine question, and not one this article will settle. But drawing the line, deciding that this is no longer your concern, is not a neutral act. It is itself a judgment. And that is where the last and most demanding form of critical thinking is needed.
Thinking against yourself
Because, in the end, critical thinking also applies to itself. As much as we think critically about AI, we can scrutinise how we think about it.
Rather than viewing everything through a negativity bias, critical thinking involves considering issues from multiple viewpoints, especially angles we are less accustomed to. Socrates, Kant, and Montaigne taught us to think “against ourselves”, challenge our own prejudices, positive and negative alike, and suspect our own conclusions. Exercising critical thinking about AI is a matter of discernment, not a radical posture, let alone a form of fanaticism.
Critical thinking must not become a respectable veil hiding conservatism, fear of change, and torpor. Properly exercised, it can have precisely the opposite effect: shaking off inertia, opening minds, and accelerating learning.
If you are an AI enthusiast, make the effort to acknowledge its risks and costs. If it repels you, now is the time to get started.
Author
Antoine HENRY de FRAHAN
Management consultant
Affiliate Professor, EDHEC Business School