Three Northern Ireland-based researchers released a brief paper with an outdated title in 2017. It was referred to as a manifesto. The majority of technology manifestos are full of exaggerated claims about disruption, scale, and the timely arrival of the future. This one had the opposite effect. It asked, quite bluntly, if the developers of the software had given any thought to the people who would have to live within it.
The authors of the paper, Raymond Bond, Jennifer Boger, and Maurice Mulvenna, were from Ulster University. It offered a set of “ethical by design” guidelines for taking ethical considerations into account when designing and implementing technology-based goods and services, and it was open about being incomplete—a work-in-progress outlining the need for fresh, creative ideas and methods in ethical design-based thinking. A group of academics acknowledging in print that they haven’t figured everything out yet is disarming. It reads more like a confession than a product launch.
One argument that continues to cause engineers to shift in their seats is what gives the document its advantage. It makes the case that design thinking ought to and is capable of being “ethical by design” and that designs ought to make an effort to transcend the moral standards established by authorities such as regulatory bodies. Beyond. Not in accordance with the past. That word works very quietly. It implies that adhering to the law is the floor, not the ceiling, and that a system may be entirely lawful while still subtly harming its users.
It’s important to consider how uncommon that position was and is, for the most part. Writing principles has been the prevailing inclination in tech ethics. Over 160 different sets of AI guidelines were once catalogued by AlgorithmWatch. Codes of conduct are released by companies in the same manner as press releases. However, the ongoing scandals reveal something that the authors of the manifesto appeared to understand early on: a wall-mounted guideline has little effect on what happens at three in the afternoon when a sprint is running behind schedule and a feature ships. Reading the surrounding scholarship gives the impression that this movement is more interested in plumbing—the unglamorous architecture underneath—than in declarations.
The manifesto was presented bluntly in the earlier work on which it is based. The objective is to maximize ethical technology design and application by going beyond meeting the currently recognized ethical principles of “autonomy, beneficence, non-maleficence, and justice.” The purposeful refusal to be exclusive is what strikes me. Regardless of their familiarity with ethical standards or the field of application, the manifesto is intended to provide people from all fields, industries, and levels of involvement with a means of participating in the discussion and making informed decisions. Not a priesthood. No gatekeeping. The junior developer and the ethicist share a seat.
The concept is beginning to permeate other areas of academia. A different group has advocated for breaking away from what they refer to as “the one best way” that the Agile Software Development process has evolved into. Instead, they suggest valuing tacit knowledge by granting the client the option to reject and to slow down or depressurize the development process. Go more slowly. That might be the most subtly radical statement anyone has made in years in a field that values speed.
It’s still unclear if any of this makes it to the typical codebase. Product roadmaps typically don’t increase citation counts. The movement is still tiny and is discussed more in whiteboard-equipped conference rooms than in those with revenue targets. However, the fundamental question it raises—not whether it’s legal, but whether it’s right—has a tendency to linger. It’s difficult to avoid thinking that the people who wrote the manifesto won’t be the ones who respond to it in the end. They will be the person who takes over the software that the rest of us are currently developing.
