The principle that a policyholder's premium should reflect the risk they actually represent, without penalising them beyond their own exposure.
Actuarial fairness is the idea that each policyholder should pay a premium proportionate to the risk they bring to the pool, neither more nor less. It justifies rate segmentation, the very foundation of insurance technique, since treating different risks alike would be inequitable toward good risks. But this principle is in tension with that of non-discrimination, which prohibits basing the price on certain protected characteristics, as illustrated by the Test-Achats ruling on gender. The rise of AI pricing models revives this debate, because very fine segmentation can track risk closely while reproducing biases or indirect discrimination. The AI Act, by classifying life and health pricing as high-risk, places precisely this trade-off between actuarial fineness and social equity under regulatory supervision.
Overpricing a policyholder solely on the statistical ground that a group they belong to reports more claims may be actuarially defensible but legally questionable in light of non-discrimination.
fairness actuarielle, actuarial fairness