The legal profession has spent the past year reassuring itself with a single, endlessly repeated slogan circulating across LinkedIn and conference panels alike.
“AI won’t replace lawyers. Lawyers who use AI will.” It sounds optimistic, practical, and actionable, giving lawyers something concrete to pursue.
The message tells lawyers to learn prompting, experiment with tools, become AI fluent, and consider themselves prepared for whatever transformation lies ahead.
But after a conversation with Dharshi Harindra, assistant general counsel for APAC at FTI Consulting, that familiar reassurance starts to feel insufficient and somewhat beside the point.
Harindra challenged the dominant narrative in a way that reframes the entire debate, moving the focus away from tools and back toward something far more fundamental.
“AI is a mirror,” Harindra said. “It reflects back at us the existing judgment that we have.”
That single observation shifts the conversation entirely, because it means AI adoption is not primarily a skills challenge but a professional reckoning.
Harindra described judgment in a precise and practical way, defining it as deciding “what matters, what doesn’t matter, what happens next, and who owns the consequences.”
Notably absent from that definition is legal research, drafting ability, or the capacity to recall every relevant case from memory under pressure.
What Harindra describes instead is decision-making under genuine uncertainty, which is the actual daily work of lawyers operating inside businesses and legal departments.
Most legal departments are not valued because their lawyers know where to find information, since information is now broadly and rapidly accessible to almost everyone.
They are valued because businesses constantly face situations where several reasonable answers exist simultaneously, none of them are perfect, and someone must recommend a path the organisation can actually execute.
Olga V. Mack, the CEO of TermScout, wrote about the conversation for Above the Law, noting that one idea kept surfacing throughout their exchange.
Mack found herself questioning whether AI had actually created far less disruption than widely assumed, and had instead exposed something that was always present within the profession.
The implication is striking, because it suggests the legal industry’s real challenge is not learning to use new tools but confronting what those tools now make visible about individual and institutional judgment.
Lawyers who have relied on process, volume, and information retrieval as their primary value proposition may find that AI does not threaten them from outside but simply removes the cover those activities once provided.
The professionals most secure in an AI-integrated legal world are therefore not necessarily those who adopted tools earliest, but those whose judgment was genuinely strong before any of it arrived.