Part of Temporal Constructs Appendix, which contains details behind the research: dates, quotations, descriptions of instruments, and sources.

General definition

The codes and frameworks that govern the use of artificial intelligence (AI) in this research, at the institution and at European level, and the points at which the practice is most exposed.

Relation to the practice

The doctorate is held to these codes, and the record of them is kept here. No rule about artificial intelligence comes from the programme itself. What is specific to it comes from the institution’s own framework and from the guidelines of the European Commission, and most of it is already in place in the practice.

Frameworks

No rule about artificial intelligence comes from the programme itself. The programme requires that the research be conducted to the ethical codes and standards of the project, which is the professional conduct learning goal. The programme’s ethics module covers that goal, and it requires a data management plan.

Two levels above the programme set what is specific to artificial intelligence. Hogeschool Inholland adopted a framework for responsible AI use in March 2026, and it replaced the institution’s earlier statements for staff and students. The framework requires that the use of such systems be open and stated. It requires that their output not be trusted in place of the researcher’s own judgement. It requires that bias and unequal effects be watched for, that the environmental cost be weighed, and that what is shared with an AI system be handled carefully. The protection of personal and sensitive data, of the institution and of third parties, follows from that requirement.

The European Commission’s living guidelines on the responsible use of generative AI in research add two requirements. Substantial use must be disclosed, with an account of what it contributed. The non-deterministic character of these systems must be allowed for in what is kept, which includes prompts, logs and model versions. The researcher remains accountable for what an agent does on their behalf.

Rules already in force

Openness is carried by two rules. The provenance that every item in the archive receives names the model and its version. Authorship keeps the artist’s own words and corrects only their fluency. The refusal to trust the output is the rule settled in July 2026. A report from an agent is not evidence until it has been compared with the machine. The cases are kept on the agent‘s page with their dates.

Question of itself

Care with data follows from where the agent runs. Its calls go through services the artist chooses, with what those services log taken into account. The practice notes that data protection is the binding constraint in Europe. The research states an ethical question for itself: the boundary between thinking with an artificial intelligence and being thought for by it. It wrote that question into the methodology and ethics text of October 2026.

Terms not yet settled

The terms on which a model trained on KML and HMSL may be used, shared or sold are still open. They are where the practice is most exposed. The data management plan that the ethics module requires sits in the programme’s paperwork. The account of who decided what, on which the demand for accountability and the Session record both turn, exists as session history. The coded decisions are what the record still lacks.

Unforeseen

Here the unforeseen situations, sometimes difficult, sometimes unexpectedly interesting, will be described as the project runs. Many of these unforeseen situations have already occurred and have been recorded. They will be added in the time to come.