AI Assurance · Harmonisation

    Overlap, delta, and the artefact that satisfies both

    Most AI obligations across major jurisdictions are the same requirement written by different hands. The work is not building a second programme — it is knowing precisely where the second rulebook diverges, and producing the extra evidence only there.

    Home jurisdiction

    Human oversight

    Home

    EU AI Act Art. 14

    Host

    SR 11-7 effective challenge

    Overlap
    Both require a person able to change the outcome.
    Delta
    US supervisors test whether challenge altered a decision; the EU tests whether oversight is designed into the system.
    Evidence artefact
    Decision log showing oversight point, reviewer and override disposition.

    Technical documentation

    Home

    EU AI Act Annex IV

    Host

    OCC 2011-12 model documentation

    Overlap
    Same artefact, different index.
    Delta
    Annex IV requires the intended purpose and foreseeable misuse; the OCC requires the theory and limitations.
    Evidence artefact
    Single model dossier with both indices generated from one source.

    Post-market monitoring

    Home

    EU AI Act Art. 72

    Host

    Ongoing monitoring (SR 11-7)

    Overlap
    Continuous performance surveillance.
    Delta
    The EU expects incident reporting to the authority; the US expects escalation to model risk management.
    Evidence artefact
    Drift monitoring output with dual escalation routing.

    Mappings shown are a published sample for European Union. The full obligation set is derived from the cross-jurisdiction AI control library and scoped to your licences under engagement.

    See the control set