Old Laws May Apply to New AI

A recently filed class action proposes that 𝗲𝘅𝗶𝘀𝘁𝗶𝗻𝗴 𝘀𝘁𝗮𝘁𝘂𝘁𝗲𝘀—not shiny new AI laws—be interpreted to apply to current AI use. A lawsuit against Eightfold AI is the first major US case to use the Fair Credit Reporting Act (FCRA) against AI hiring software. The plaintiffs in 𝘌𝘪𝘨𝘩𝘵𝘧𝘰𝘭𝘥 allege that the Eightfold AI hiring platform violates the FCRA by generating hidden “match scores” that function like consumer reports without giving applicants notice or the ability to dispute errors.


𝗣𝗹𝗮𝗶𝗻𝘁𝗶𝗳𝗳'𝘀 𝗹𝗲𝗴𝗮𝗹 𝘁𝗵𝗲𝗼𝗿𝘆 𝗿𝗲𝗳𝗿𝗮𝗺𝗲𝘀 𝗔𝗜‑𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲𝗱 𝗰𝗮𝗻𝗱𝗶𝗱𝗮𝘁𝗲 𝗿𝗮𝗻𝗸𝗶𝗻𝗴𝘀 𝗮𝘀 𝗿𝗲𝗴𝘂𝗹𝗮𝘁𝗲𝗱 𝗰𝗿𝗲𝗱𝗶𝘁‑𝘀𝘁𝘆𝗹𝗲 𝗮𝘀𝘀𝗲𝘀𝘀𝗺𝗲𝗻𝘁𝘀—𝗮 𝗻𝗼𝘃𝗲𝗹 𝗮𝗻𝗱 𝗽𝗼𝘁𝗲𝗻𝘁𝗶𝗮𝗹𝗹𝘆 𝘀𝘄𝗲𝗲𝗽𝗶𝗻𝗴 𝗶𝗻𝘁𝗲𝗿𝗽𝗿𝗲𝘁𝗮𝘁𝗶𝗼𝗻.

𝟭. 𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀: If the court agrees, any AI vendor that creates applicant scores, rankings, profiles, or predictions may be treated as a consumer reporting agency—subject to strict federal rules. And, this case directly challenges the black-box nature of AI algorithms in handling human profiles: plaintiffs argue they had no idea Eightfold was generating dossiers and scoring them from 0–5 based on inferred traits like behavior, aptitude, and personality.

𝟮. 𝗣𝗼𝘁𝗲𝗻𝘁𝗶𝗮𝗹 𝗔𝗳𝘁𝗲𝗿𝗺𝗮𝘁𝗵: Courts could force algorithmic hiring systems to become auditable, explainable, and correctable.

𝟯. 𝗗𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗹𝗲𝗴𝗮𝗹 𝗮𝗻𝗴𝗹𝗲: Unlike other lawsuits, this one focuses on failure to provide notice, transparency, and an opportunity to challenge errors. Such procedural rights are often easier to prove than discrimination.

𝗣𝗼𝘁𝗲𝗻𝘁𝗶𝗮𝗹 𝗴𝗮𝘁𝗲𝘄𝗮𝘆 𝗰𝗮𝘀𝗲: Eightfold may open the door for courts to see existing legal regimes as equally applicable to new AI practices. Two time periods, one law—with the former rules transforming to cover new technology.