This essay is part of Ways of working for more peaceful futures: A global health humanities roundtable, an effort through the Future of Peace program to bring perspectives from global health humanities into conversation with questions of peace. Read Michelle E. Anderson’s introduction to the series, or explore the full roundtable paper bringing all six contributions together.
Metacognitive competence
The global health challenges I have encountered—from developing education programs, devising public health interventions, and creating transfusion and cellular-therapy capacity through to working across professional and ethnographic cultures and unequal health systems— have rarely been solved just by knowledge acquisition and cumulative information. They demand another underrecognized domain of cognitive skills and competencies known as metacognition: the awareness and learned capacity to notice, question, and regulate one’s own thinking, emotions, assumptions, and biases. This field has in recent years been deeply enriched by findings in neuroscience and entails a range of other competencies, such as how we perceive, reflect, navigate change, pay attention, make decisions, and deal with mistakes.
Practitioners of medicine and global health often work and act under conditions of uncertainty, time pressure, and moral urgency. Those conditions tend to reward knowledge, decisiveness, and authority. But these attributes can lead to overconfidence, biased decision-making, and premature conclusions. The assumption is that technical know-how and knowledge acquisition confer wisdom and expertise. Metacognition creates a pause: What am I assuming? What, and whose, knowledge is absent? What emotion is shaping this judgment? What evidence would change my mind?
From traditional expertise to learning and adaptive expertise
My work in medical training and education, transfusion medicine, and global professional collaborations has taught me that technical competence is important and often essential but not always sufficient. A protocol that is appropriate in a well-resourced cancer center may be unsafe, impractical, or inequitable elsewhere—not because colleagues lack knowledge, but because contextual realities differ – the local pathologies, cultures, political conditions, infrastructures, and patients’ living conditions and realities.
A critical foundation for developing metacognitive competence is epistemic humility: recognizing that one’s expertise is always partial. It means entering partnerships not simply asking, “What can we teach (given our expertise)?” but “What do we need to understand before acting?” It requires treating local clinicians, patients, and communities as co-interpreters of a problem, rather than as passive recipients of externally created solutions.
This approach is especially important in global health, where even good intentions can lead to acceptance and reproduction of unequal power structures. Programs framed as “capacity building” may overlook existing and reliable capabilities; data-driven priorities may displace legitimate community priorities; and urgent interventions may neglect the long-term relationships required for trust. Metacognitive awareness does not slow action for its own sake. Instead, it ensures that action is not merely efficient but responsive and adaptive. This approach leads one from traditional expertise based on knowledge acquisition, to adaptive expertise based on adaptive cognitive processes and how knowledge is applied.
Thinking adaptively and collectively
Metacognition is not only an individual practice but can also be a collective practice guiding a team’s ability to examine how it reaches conclusions, whose voices carry authority, and when consensus may conceal dissent. In multidisciplinary clinical teams, this may suggest inviting the person closest to the patient or operational problem to challenge an apparently expert decision. In international partnerships, it means building structures in which reciprocity is authentic, and disagreement is safe.
Such shared cognition is a practical form of peacebuilding. Violence is not limited to war. It can also be embedded in disregard, extraction, exclusion, and systems that make some lives easier to overlook than others. Institutions become less violent when they can detect their own blind spots, acknowledge harms, and revise their practices with inbuilt safety and without defensiveness.
In working towards peace and just futures, listening must be active and attentive, not interrupted or feigned for politeness. It must have consequences. If communities identify a problem differently from external experts, the response should not be to explain their view away, but to ask what the difference reveals about the limits of one’s own frame.
AI, human judgment and metacognition
Artificial intelligence intensifies the need for this discipline. AI can retrieve and synthesize enormous quantities of information, recognize patterns, generate plausible recommendations, and support work in settings where expertise is scarce. These capacities could be valuable in global health in analyzing data, supporting diagnostic decisions, interpreting materials, widening access to educational resources.
But AI can also create a misleading illusion of neutral, slick knowledge. Its outputs reflect the data, institutions, and inequalities from which it is built. A system trained mainly on data from affluent settings may perform poorly in populations, languages, or health systems that were inadequately represented. Its confident language can conceal uncertainty. And its ostensible objectivity can encourage users to defer to it rather than to the people affected by its recommendations.
The essential skill, therefore, is not merely AI literacy—knowing what AI tools can do—but meta-AI literacy: the metacognitive skill to recognize how AI tools shape our judgment and the metacognitive competencies to deal with the range of threats posed by AI. When an algorithm offers a diagnosis, priority, or prediction, we should ask: What is this system seeing? What is it unable to see? What assumptions are embedded in its categories? Why am I inclined to trust or reject it? Who bears the cost if it is wrong?
AI should complement if not extend human capacities but not assume its own ethics and replace human moral responsibility. A machine can rank options, but it cannot decide what justice requires. An AI tool can identify correlations that also result in reckless policies. While it can produce answers, it cannot participate in the deeply human work of repairing trust.
Metacognition will not eliminate error, conflict, or injustice. Its value lies in making corrections attainable and guiding us there. In a world of accelerating technology, polarized politics, and deep global inequality, the capacity to examine how we know—and how we may be wrong—is not a luxury but a condition for acting with humility, solidarity, and peace.
