Moral imagination

A peer-reviewed method, facilitated with the team that controls the tech.

---

Ethics review boards and compliance processes work well once a product exists. But they often reach teams too late for the choices that matter during ideation and prototyping, while the team is codifying their intentions.

We developed and led this method at Google over several years, working directly with engineering, research, and product teams, and refined it across more than seventy sessions. The method has been peer-reviewed across several papers.

Moral imagination brings ethical thining into the design process. We work with a team while its options are still open, using interactive and engaging exercises to surface the values already in play, challenge the assumptions, and turn the result into commitments the team writes into its own planning documents. We don't lecture on ethics or import a framework from outside, but teams only act on conclusions they reach themselves with our expert guidance.

How it works

A workshop takes the current technical project your team already works on and deliberates on it towards an ethical conclusion the team can defend. There are four steps. What goes into each one is built around your context, but the order doesn't change, as it's based on rigorous ethical methods

Reflection: The team says out loud what it already believes. Semi-structured discussion surfaces the values people bring to the work, the benefits they expect the technology to deliver, and the vision behind it. Teams routinely find a gap between the values they state and the practices they follow. The real work starts once those implicit intentions are made explicit and understood.

Expansion: We push back on the picture the team just made visible. A bespoke techno-moral scenario projects your technology five to ten years forward and stops on a cliffhanger, forcing the team to anticipate rather than judge in hindsight. Role-play, a veil-of-ignorance exercise, and a taxonomy of sociotechnical harms bring in the stakeholders missing from the room and the failure modes that were not previously considered.

Evaluation: The team reasons through the tensions it surfaced. We introduce moral theories as tools, not doctrine, along with the distinctions that makes deliberation possible. We weigh competing values and principles, working towards resolutions. Participants argue positions that aren't their own and work out where the disagreements lie.

Action: The surfaced insights then become commitments. We turn the ethical focus areas into responsibility objectives written in the language your organization already uses, so they go straight into PRDs and OKRs. Objectives the team negotiated provide more legitimacy and durability than imposed checklist.

What teams leave with

Teams leave with a decision reasoned backwards from the world they want to build, rather than forwards from whatever the technology happened to make possible. That reasoning is documented, so it can be defended later to users, a regulator, a business partner, or a journalist.

They also leave with responsibility objectives, written in their own planning language and owned by the team that wrote them. These go into PRDs and OKRs, and give the team standing when it deals with review boards and other formal controls.

The most durable output is a shared vocabulary. Teams learn to recognize when a technical conversation has turned into one about values, and can run that conversation without moral technologists in the room.