AI systems may be subject to specific applicable legal and regulatoryrequirements. Some legal requirements can mandate (e.g., nondiscrimination,data privacy and security controls) documentation, disclosure, and increased AIsystem transparency. These requirements are complex and may not beapplicable or differ across applications and contexts.
For example, AI system testing processes for bias measurement, such asdisparate impact, are not applied uniformly within the legal context. Disparateimpact is broadly defined as a facially neutral policy or practice thatdisproportionately harms a group based on a protected trait. Notably, somemodeling algorithms or debiasing techniques that rely on demographicinformation, could also come into tension with legal prohibitions on disparatetreatment (i.e., intentional discrimination).
Additionally, some intended users of AI systems may not have consistent orreliable access to fundamental internet technologies (a phenomenon widelydescribed as the “digital divide”) or may experience difficulties interacting withAI systems due to disabilities or impairments. Such factors may mean differentcommunities experience bias or other negative impacts when trying to accessAI systems. Failure to address such design issues may pose legal risks, forexample in employment related activities affecting persons with disabilities.
Views: 0
