How should contextual adequacy be defined without turning it into surveillance?
Any approach must distinguish interactional context from claims about private inner life.
Open questions
What matters is not pretending that we already have every answer.
A programme of research is also a public account of the questions it has not resolved.
Which forms of evidence are useful for review, and which create disproportionate retention?
Traceability needs boundaries, proportionality and a clear purpose.
When is human intervention required, meaningful and possible?
Human oversight should be an operational condition, not a decorative label.
How can an empirical study remain genuinely contestable?
Protocols, limitations and independent scrutiny must be part of the work from the beginning.