Wa
Most approaches to AI governance follow the tested route of law or policy and therefore centre a rules-based mechanism that calls for compliance and conformity, and punitive consequences for non-conformity. This is part of a larger ethos that is cyclical: Where regulatory bodies set rules, companies contest them, advocates push back, and the framework that emerges tends to most align with the power balance that defined the negotiation process. As a result, governance that emerges is legalistic, minimally compliant, and regularly contested. Pushing back against this contested terrain, the Japanese practice of Wa, translated simply to mean harmony, offers another route to govern AI.
Before proceeding, we want to name the fact that Wa runs deep in Japanese culture, shaping everything from workplace dynamics to political life. We approach it as learners, with the awareness that its inclusion in this archive carries a tension that we do not want to smooth over: The same ethic of harmony that fosters genuine collective care has been used to suppress dissent, silence minorities, and maintain unjust social arrangements in the name of group cohesion. We hold this tension openly throughout this entry because it is precisely where this concept becomes most useful for AI governance.
Understanding Wa
Wa (和) is often translated to mean harmony. However, in practice and its embodied presentation, it points to something more specific, namely the active sustenance and maintenance of smooth, non-conflictual relationships within a group. Wa, in its true form, is not about quelling dissent or expecting conformity, but rather also shapes the mode of expressing disagreement in principled ways that prioritise the relationship rather than cause rupture. In Japanese cultural life, Wa shapes the ways in which a decision is made, how individual desires are weighed against collective needs, and how disagreement is acknowledged as generative rather than silenced.
Wa does not seek the absence of conflict or offer a negative connotation to peace. It aligns with the idea of positive peace, where there is an active, ongoing, and effortful cultivation of a social environment in which people can work through harm and coexist without rupture. In this sense, Wa is a genuine ethic of collective care that includes a recognition of an individual’s wellbeing as integrated in the community’s wellbeing. It calls on recognizing the importance of listening before speaking, seeking a consensus rather than imposing on another, and to maintaining a relationship even through the difficult times. However, it is also important to acknowledge that Wa could impose a sense of social pressure that silences those who may name harm or call for the protection of collective reputation or values at the cost of an individual’s truth or voice. It can also demand that those with the least amount of power to absorb the discomfort of arrangements that serve those with the most power. This inherently makes Wa a powerful principle that can be genuinely instructive for governance, because it presents a tension that governance in practice in itself can produce and therefore must confront.
Wa informs the ethical approach of Nemawashi, or the slow and patient process of building a consensus by consulting all affected parties before a decision is made, and Ringi, or the practice of circulating proposals for input before they are finalised.
What Wa unsettles in AI governance
Wa counters the compliance-first approach to AI governance and asks if a different process could produce different outcomes that centre consensus-seeking, meaningful consultation, and a sustained commitment to maintaining relationships that make ongoing governance possible.
Specifically, Nemawishi, which is informed by Wa, asks us to think about making decisions after consultations rather than the other way around. In AI governance, this would mean that communities that are most exposed to AI harms being involved in designing governance frameworks from the inception, rather than be forced to participate in a consultation where their inputs are sought on frameworks already drafted by others.
Wa also has a shadow that is equally illuminating. While AI governance presents harmony in the form of multistakeholder processes that include civil society representatives but offer them no real power, impact assessments that may identify harms but create no obligation to address them, and ethical review boards that protect institutional reputation instead of the people at risk, this harmony is superficial and performative. These are the governance equivalents of demanding Wa while refusing to address the underlying conflict. Wa’s own internal critique of the tensions inherent in collective processes holds up a mirror to these realities within approaches to governance in their current form.
Wa asks us to pay attention to the kind of harmony we are actually maintaining in AI governance and whose discomfort is being absorbed so that others can experience smoothness. It tells us that genuine and earnest harmony requires the inclusion of everyone who is present, in substantial terms. It also asks us to guard against false harmony that causes certain voices to be suppressed. The question is not whether to pursue collective cohesion, but whether that cohesion in itself is built on honesty or the silence of the powerless and dominance of the powerful.
References
Wierzbicka, A. (1991). Japanese key words and core cultural values1. Language in society, 20(3), 333-385.
Lee, O. (1983). THE CULTURE OF WA. Japan Quarterly, 30(1), 54.
Triplett, K. (2006). The discourse on wa or harmony in contemporary Japanese religions and society. Religious harmony problems, practice, and education, 179-187.
Zhao, M. Y. (2026). Harmony: towards a unified conceptual framework. The journal of positive psychology, 21(2), 241-254.
