Institutional Answerability
How responsibility is distributed when institutions delegate judgement to predictive systems.
Theory
Institutional answerability extends the concept of answerability to the collective agents — hospitals, courts, universities, agencies — that increasingly act through AI-mediated processes. It asks how such bodies remain accountable when constitutive judgements have been delegated.
The account rejects both the fiction of a single 'responsible engineer' and the diffusion of accountability into 'the system.' Instead, it identifies specific institutional roles, procedures, and records through which answerability can be preserved or lost.
The framework offers institutions a diagnostic: which of our answerability relations have been left implicit, which have been quietly transferred, and which have been abolished by the introduction of predictive tooling?
Key publications
- Institutional answerability and algorithmic governance (AI and Ethics).
- Case study of answerability failures in predictive administration (Behavioral Sciences).
- Theological reflections on institutions as moral agents (Studies in Christian Ethics; New Blackfriars).
A continually updated record is available through the publications page.
Current work
- A framework for institutional answerability audits of AI deployments.
- Ongoing collaboration with public-sector partners on governance of predictive systems.
- Contribution to a volume on collective responsibility in algorithmic societies.