J. Paul is a London based designer and researcher with expertise in Speculative Design, Service Design, Design Research, and Strategy.
NKD_3 — Good is good and bad, and bad is bad and good, and often we can't tell the difference.
This is the principle most likely to be misread. It is not a license for moral relativism. It is the opposite: a discipline against the premature moral certainty that lets badly-thought-through interventions ship under the banner of obvious good.
The misconception it corrects
The dominant mode of corporate ethics is clean sorting. Initiatives are coded as good (sustainability, accessibility, mental health, community) or bad (carbon, exclusion, addiction, surveillance). The good ones get accelerated; the bad ones get mitigated or hidden; everyone goes home feeling that the moral work has been done.
The world does not sort cleanly. Sustainability initiatives that increase emissions through Jevons-effect rebound. Mental health products that pathologise normal difficulty. Community features that produce mobs. Accessibility tools that strip dignity. The list of well-intentioned interventions that produced their opposite is the standard reading list of any honest design ethics class.
NKD_3 names the phenomenon: that the moral valence of an outcome is rarely settled by the moral framing of the intent. Good and bad are usually entangled in the same outcome, and the most dangerous moment in a project is the moment the team decides which one it is.
What it changes in practice
Goodness is held as a hypothesis, not a label. When a team frames an initiative as "good," that framing is treated as a claim about the world that the team is obliged to test. What evidence would falsify the claim? Under what conditions does the good outcome flip? NKD_3 turns ethical framing into a research question.
The opposite outcome is designed for. For every intended good, the team explicitly imagines the configuration of users, contexts, or feedback loops that would produce the inverse. Not as risk management — as a design input. The opposite outcome is built into the brief.
Moral framings get expiry dates. Initiatives marked "good" at launch are revisited at fixed intervals with the question are we still confident this is good? The answer is sometimes no. Teams that take NKD_3 seriously become much better at killing their own "good" projects when the evidence turns.
A working method: the inversion brief
The single technique we use most often for NKD_3 is the inversion brief. It works as follows.
For any project framed in moral terms — we are designing for sustainability, for accessibility, for wellbeing, for community — write a one-page brief for the inverted project: we are designing for unsustainability, for exclusion, for ill-being, for atomisation.
The brief should be specific: what would the product look like, what users would it most impact, what feedback loops would amplify it, what business case would justify it.
Two things happen.
First, the team almost always discovers that some elements of the inverted brief are already present in the good brief. The good initiative, on inspection, contains plausible mechanisms for the inverse outcome. Surfacing these early is enormously cheaper than discovering them in user research two years later.
Second, the team gains a calibration tool. The intended-good brief and the inverted brief sit side by side, and the question becomes: which of these is the actual likely outcome, given the system we are putting this into? That is a different question from which one do we intend?, and a much more useful one.
A note on what this principle is not
NKD_3 is not "everything is grey" relativism. It does not say that moral judgement is impossible or that all outcomes are equivalent. It says that premature moral judgement is dangerous, and that the work of design is to keep moral assessment open under conditions of irreducible ambiguity for as long as possible — long enough to design under reality rather than under flattery.
It also does not say that designers should not have values. Quite the opposite. Designers who take NKD_3 seriously tend, in our experience, to have stronger and more durable values than those who do not, because their values have been tested against the inversion rather than asserted alongside it.
Where this principle lives in our work
NKD_3 is the principle that shows up most in our work in healthcare and technology — two domains where the gap between intended-good and actual-good is widest, and where the cost of getting it wrong falls on the people most poorly placed to defend themselves. We do not run engagements in these domains without the inversion brief.
Key takeaway: moral framing is a hypothesis, not a label. Design the inverse of the intended outcome alongside the intended one, and use the comparison to ask which is the actual likely result.