AI Exposes How Our Definition of Competence Has Shifted Away From Judgment

Every era believes it grasps the essence of its most vital words, even as the realities they represent subtly evolve. Words like “freedom,” “responsibility,” “authority,” and “trust” have endured centuries of change, yet their meanings are not precisely what they were for past generations. “Competence” may have undergone a similar metamorphosis, though it’s particularly hard to detect because we continue to use the term with unwavering confidence. We readily label physicians, judges, teachers, engineers, and leaders as “competent,” as if the quality were self-evident and unchanging. However, the very characteristics used to identify competence have shifted significantly, and AI may now be highlighting the extent of this transformation.

For a considerable portion of professional history, competence was most evident when certainty was unattainable. The skilled physician possessed medical knowledge, but the challenging patient was precisely the one for whom established practices and evidence offered no clear-cut solution. The adept judge understood the law, yet judgment was crucial where applying a general legal principle to a specific human circumstance demanded interpretation. The effective leader grasped the organization’s workings, but leadership truly mattered when precedent could no longer dictate the path forward. Thus, competence inherently involved the capacity to act responsibly when existing rules, knowledge, and experience reached their limits.

Knowledge formed the bedrock upon which judgment rested. Yet, professional competence also entailed an acknowledgment that reality could occasionally transcend the systems designed to encompass it. Expertise was valuable not because the expert always had an answer, but because the expert could recognize when an answer remained uncertain and still take ownership of the next steps. Modern institutions had valid reasons to feel uneasy with such an ambiguous definition. As organizations expanded, professions became more specialized, and decisions carried greater weight, institutions needed to identify, certify, cultivate, and evaluate competence on a large scale. Medical education introduced standardized tests and professional guidelines; corporations developed performance metrics and competency frameworks; universities formalized assessment procedures; and regulators mandated demonstrable standards. These developments represented genuine progress, reducing arbitrary authority, enhancing the transparency of professional standards, and enabling expertise to disseminate beyond its original developers.

However, a subtle shift occurred. The indicators used to recognize competence gradually became conflated with competence itself. Consistency began to signify reliability, procedural accuracy to represent professionalism, measurable performance to denote effectiveness, and confident communication to imply understanding. The problem arises when the very characteristics designed to make competence visible become its definition. This distinction is critical because what can be measured and standardized tends to be what institutions excel at recognizing. A professional who delivers clear answers, adheres to established processes, communicates fluently, and performs consistently leaves a substantial trail of evidence of their competence. Conversely, an individual whose primary contribution lies in recognizing that the existing framework is no longer suitable for the situation may leave far less evidence. The more organizations rely on observable signals to identify expertise, the greater the risk of prioritizing the appearance of competence over its actual practice.

This transformation began long before the advent of AI, which is significant because it exposes a problem that technology did not create. Generative AI systems can now produce many of the signals that modern institutions have come to associate with competence: articulate language, structured reasoning, rapid synthesis, polished presentations, confident recommendations, and plausible explanations. Tasks that once demanded considerable professional effort can now be accomplished in seconds. The immediate conclusion is that AI threatens traditional forms of expertise. A more profound possibility is that AI is revealing how limited our definition of expertise had already become. If eloquence can be generated, consistency automated, information accessed instantly, and plausible answers produced at minimal cost, then perhaps these qualities were never competence itself. We may be entering an era where the visible performance of competence becomes increasingly common, while cultivating genuine competence remains as challenging as ever. A well-structured analysis might reflect considerable judgment or simply the proficient use of a machine. The output alone increasingly tells us less about the intellectual process behind it.

The constructive approach is not to retreat into an elusive realm of human intuition inaccessible to machines, but to identify which dimensions of competence become more crucial as its traditional signals grow less informative. Here, the older understanding proves unexpectedly relevant, reminding us that competence was never solely about producing correct answers. Competence also involves discerning which questions are worth asking, differentiating the relevant from the merely available, understanding when a general rule applies to a specific situation and when it does not, identifying uncertainty rather than concealing it, and accepting responsibility for outcomes that cannot be delegated to the informing system. None of these qualities necessitates rejecting AI. Intelligent systems can enhance each of them by broadening the information available to professionals, challenging assumptions, and revealing patterns that humans might overlook. However, better information does not negate the need to interpret its meaning within a specific context.

This has implications for how institutions educate, recruit, and evaluate individuals. Universities might need to focus less on students' ability to reproduce convincing answers and more on their capacity to question them. Organizations may need to look beyond fluency and responsiveness when identifying talent. Professional education might need to emphasize uncertainty, interpretation, and responsibility more explicitly. Leadership development could benefit from placing greater value on recognizing when an established framework no longer reflects the reality it was designed to manage.

None of this requires abandoning standards, metrics, or procedures. Modern institutions rely on them to function, and reverting to purely subjective assessments of expertise would reintroduce the very problems standardization aimed to solve. The challenge lies in retaining the advantages of making competence visible without mistaking visibility for competence itself. Perhaps, then, AI is not fundamentally altering the meaning of competence but rather compelling us to acknowledge that we had already redefined it. Over decades, our understanding shifted from one centered on judgment under uncertainty to one increasingly associated with consistency, demonstrable performance, and the reliable generation of answers. This transition yielded immense benefits but also narrowed our perspective. Now that machines can replicate many of these visible characteristics with remarkable ease, the distinction we allowed to fade has become impossible to ignore. The crucial question is no longer whether AI will diminish human competence, but whether institutions will continue to recognize competence through signals that technology has rendered increasingly unreliable, or whether we will reclaim a richer understanding of what competence was always meant to embody: not merely knowing the answer, but knowing what to do when the answer is no longer sufficient.