ABSTRACT
While there has already been much discussion among lawyers about liability for artificial intelligence (AI), comparatively little attention has been devoted to the actual basis of such liability. It is often taken for granted that product liability will apply to AI, but such liability is based in all legal systems on the product’s defect, and, as the article purports to demonstrate, proving an algorithm’s defect can be a daunting, if not impossible task for plaintiffs. Other liability bases should therefore be sought, which do not make compensation for harm caused by AI illusory, while preserving the legitimate interests of those who design or operate AI. As the article suggests, there is no one-size-fits-all in this respect and different liability regimes should apply, each with its own basis, depending on the field in which AI is being used. Harm caused by autonomous vehicles, for example, is probably better addressed through strict liability regimes for traffic accidents that through a general liability for AI regime.
Borghetti, Jean-Sébastien, Civil Liability for Artificial Intelligence: What Should its Basis Be? (June 1, 2019). La Revue des Juristes de Sciences Po, juin 2019, n°17 ISSN 2111-4293, 94-102.
First posted 2020-03-19 09:07:23
Leave a Reply