Skip to main content

Systemic-risk GPAI

The subset of the most capable general-purpose AI models that, when so designated, trigger heavier provider duties (state-of-the-art evaluation including adversarial testing, systemic-risk assessment and mitigation, serious-incident reporting, stronger cybersecurity). Free and open-source models get lighter GPAI obligations except when designated systemic-risk.

Defined in 3 GAGE programs, which carry 3 distinct definitions of it. The wording above is taught in AI Literacy & Professional Conduct.

How each discipline defines it

The same term does different work depending on who is using it. These are the definitions as each program teaches them, unedited.

AI Literacy & Professional Conduct

The subset of the most capable general-purpose AI models that, when so designated, trigger heavier provider duties (state-of-the-art evaluation including adversarial testing, systemic-risk assessment and mitigation, serious-incident reporting, stronger cybersecurity). Free and open-source models get lighter GPAI obligations except when designated systemic-risk.

The AI Lobbyist: Elite AI Policy Influence Program

the heavier EU AI Act obligation tier (Article 55) for general-purpose AI models presumed high-impact because their cumulative training compute exceeds 10^25 floating-point operations (Article 51), or because the Commission separately designates them as systemic-risk under Annex XIII criteria.

EU AI Act Implementation Expert

A GPAI model that is presumed to have high-impact capabilities because it was trained using cumulative compute greater than 10^25 floating-point operations (Article 51(2)). Providers of systemic-risk GPAI models have additional obligations under Article 55 beyond Article 53.

Where it is taught

The exact lessons this term appears in. The first 7 topics of every program are free with a free account.

Terms it appears with

Not an alphabetical neighbourhood: these are the terms taught in the same lessons, ranked by how often they appear together.