Under the EU AI Act, a GPAI (General-Purpose AI) model is classified as having 'systemic risk' if it meets which criterion?
Select an answer to reveal the explanation.
Short Explanation
Here's the deal — under Article 51 of the EU AI Act, you presume a GPAI model has systemic risk when its cumulative training compute exceeds 10^25 FLOPs. That's the regulation's main quantitative threshold.
Full Explanation
The correct answer is the compute-threshold option: under Article 51 of the EU AI Act, a GPAI model is presumed to have systemic risk if the cumulative amount of compute used for its training exceeds 10^25 floating point operations (FLOPs). This is the primary quantitative threshold established in the regulation. The high-risk-sector option is wrong because deployment in a high-risk sector such as healthcare or financial services determines high-risk AI classification, not GPAI systemic risk. The download-count option is wrong because being downloaded more than one million times by EU-based users is not a GPAI systemic risk criterion. The synthetic-media option is wrong because the capability to generate synthetic media that could mislead the public may be a relevant consideration but is not the systemic risk threshold.