I have been working in AI training for legal departments since what now feels like a different era, long before generative models became mainstream. Today I teach AI for business and support organizations not only in selecting tools, but in implementing them at scale. Over the past few years I have observed a predictable range of reactions: fear of not knowing enough, enthusiasm for new possibilities, and a certain ego-driven confidence that a short course and a few hours on ChatGPT might be enough to master prompting and quickly turn that knowledge into competitive advantage. None of this was surprising. Every technological shift produces excitement, insecurity, and overconfidence in similar proportions.
What I did not fully anticipate, especially among in-house counsels, is something more subtle. Beneath the visible experimentation and training, there is a quieter concern that is rarely articulated openly. It is not about learning how to use the tools. It is about remaining relevant.
When I speak with them, I often sense a silent equation running in the background. In a law firm you are a fee earner. You generate revenue. Your time is billed and your value is directly tied to income. In-house, the logic is different. You are not a profit center. You are a pure cost. However strategic your role may be, however sophisticated your advice, you sit on the expense side of the balance sheet. That structural difference becomes critical when artificial intelligence changes the perceived weight of that cost.
These are not abstract anxieties. They are rational calculations. In a revenue-generating role, efficiency improves margins. In a cost center, efficiency can reduce headcount. That is the uncomfortable asymmetry in-house lawyers are now confronting.
The concern is strongest in roles below the top tier. General Counsel positions are strategic and political by nature, and organizations invest heavily in them because they are seen as a competitive advantage. The layers beneath are more exposed. Resistance, however, rarely appears as open opposition. It manifests as cautious experimentation that never fully scales, as disproportionate focus on risk without corresponding solution-building, and as compliance used as a brake rather than as an enabling framework.
Organizations face a structural paradox. They must accelerate transformation, efficiency, and integration, actively promoting widespread AI adoption. At the same time, they are expected to reassure professionals that they remain essential, even as an unprecedented productivity shift reshapes operating models. These messages sit in tension. If output per person rises materially, the arithmetic changes whether anyone acknowledges it or not. Leaders understand this. Employees understand it too, even when it remains unspoken.
The real issue is not technological adoption. That trajectory is clear. The real issue is how professional value is redefined. If value is anchored primarily in execution, information retrieval, and production speed, AI will inevitably feel like a threat. If value shifts toward judgment, strategic framing, risk calibration, contextual interpretation, and the ability to translate legal complexity into business consequences, AI becomes leverage rather than replacement.
There is another reality that is rarely voiced. Not everyone is equally competent, and not every legal function operates at a high strategic level. Some professionals bring judgment, contextual intelligence, and decision-making depth. Others primarily execute processes, replicate templates, and manage workflows that are increasingly systematizable. The arrival of AI exposes inefficiency, superficial expertise, and roles sustained more by structure than by differentiated value.
The earlier fear surrounding AI was about competence: not knowing how to use it. The emerging fear is more existential: mastering it and still wondering whether your role will contract. Ignoring that fear will not dissolve it. It will simply turn into hesitation and subtle resistance, weakening both individuals and organizations.
In the past, differentiation was largely about seniority.
Today, it is about replaceability
