Standards and expertise
Participation in technical committees, expert review of draft regulation and methodological documents that turn research into something a regulator or an engineer can use.
Where we contribute
National standardisation
The Institute is preparing an application to join the Russian technical committee for artificial intelligence (TK 164), where national standards in the AI series are developed. The application has not been filed yet; we will publish the status once it changes.
Expert review of draft acts
We plan to submit written expert opinions on draft normative acts through the official public consultation procedure, and to publish them alongside the submission.
Methodological documents
Internal and public methodology: how to scope a behavioural audit, how to define acceptance criteria for an AI system, how to record an incident so that it can later be counted.
Corporate standards
For customers who need an internal rule rather than a research paper: a document written to fit the customer's processes, with review points and a defined owner.
Themes we hold a position on
Below is our public position on these subjects. It is based on open sources; we are ready to discuss it publicly and to refine it as practice develops.
| Theme | Our position in one line |
|---|---|
| Regulation of general-purpose models | Obligations should attach to measurable behaviour and to the ability to investigate it, not to model size alone. |
| Labelling of generated content | Label the artefact and preserve provenance; a label without provenance cannot be verified. |
| Disclosure of incidents | An incident that cannot be counted cannot be governed; a common recording format is a precondition for any serious oversight. |
| Agent autonomy | Authority granted to an agent should be bounded, logged and revocable, and the log should be the object of audit. |
| Evaluation of safety claims | A claim of safety is only meaningful together with the scenario set it was tested against. |
| Publication of findings | Findings in the public interest; attack payloads only under responsible disclosure. |
Expert products
Expert opinion
A written assessment of a draft act, a technical specification or an architecture, with the reasoning shown and the limits of the assessment stated.
Threat model
Assets, actors, attack paths and controls for a specific AI system, mapped to the customer's own processes.
Methodology
A repeatable procedure: steps, roles, artefacts, acceptance criteria and a checklist that can be audited by a third party.