liveBook · Manning

Instructions for use

As explained in AI Governance

Provider-supplied directions describing a system's intended purpose, limitations, requirements, and conditions for use.

Instructions for use are provider-supplied directions for running a system safely within the bounds it has been tested for. If a deployer cannot set up monitoring from the directions alone, they are legally insufficient. section 7.3.3

For high-risk systems, the directions must explain:

  • The intended purpose and system limits. section 7.3.3
  • Required input quality and how accuracy was measured. section 7.3.3
  • The human oversight the deployer must implement and known failure scenarios. section 7.3.3

A deployer using a finished high-risk product is owed instructions for use, not the provider’s full technical file, which is intended for regulators. Treat these as distinct documentation deliverables. section 7.3.5

How AI Governance covers it

How deep it goes
Specifies purpose, limits, input quality, accuracy, oversight, and failure scenarios, while remaining distinct from the regulator-facing technical file.
What you can do after
Operate/configure it: Enables deployers to run a system within tested bounds and establish monitoring from the provided directions.
How it connects
Connects safe operation to input quality, accuracy measurement, human oversight, and monitoring.

This explainer was generated by Manning's knowledge graph from the text of AI Governance. The section links point to the parts of the book it draws on.