Anticipate consequences
What foreseeable human harms could result from a system's normal operation, errors, or misuse?
ADOC · DDOC · Prevention
Duty of Care by Design
A system should not merely accomplish its assigned task. It should also take account of what that task—and its failures—may do to human lives.
From reaction to prevention
The book proposes Algorithmic Duty of Care (ADOC) and Data Duty of Care (DDOC) as ways to extend the familiar responsibility to avoid foreseeable harm into algorithms, data, agents, and other digital systems. Its focus is broader than AI alone.
Many systems belong to what Mason and Hall call the Oblivatronix generation: designed to “just do the job” while remaining oblivious to human vulnerability. Their alternative is to build responsibility into design, code, and institutional practice.
Design is faster than legislation. We do not have to wait for the next disaster to begin building safeguards.
What could care built into a system look like?
The following are illustrative implementation questions consistent with the book's argument, not a checklist reproduced from its pages.
Connect complaints
Can the system recognize similar reports across users and locations instead of treating each person as an isolated exception?
Preserve evidence
Are warnings, decisions, changes, and overrides recorded in a way that supports meaningful review?
Trigger response
Do serious or repeated warning signs lead to investigation, escalation, and corrective action?
Make correction possible
Can harmful operations be challenged, reviewed, changed, or paused when credible risks emerge?
Keep humans accountable
Who is responsible for acting on evidence of harm, and how is that responsibility made visible?
Technical care and human care belong together
ADOC is not offered as a substitute for courts, regulators, professional ethics, or institutional accountability. The authors explicitly describe it as one part of a combined approach: technical design, law, human oversight, and other safeguards working together.
They also connect this design agenda to Artificial Compassion Technology and developments in human and medical science. The larger research program asks designers to understand the biological and social needs of people who co-inhabit the world with technology.