Computers make mistakes and AI will make things worse — the law must recognize that

Posted under: AI technologies
Date: 2024-01-25
Computers make mistakes and AI will make things worse — the law must recognize that

Over 20 years ago, Fujitsu developed the Horizon accounting software for the UK Post Office, marking Europe's largest civilian IT project rollout. However, severe errors in the system led to false accusations against Post Office workers for theft or false accounting. More than 700 people were prosecuted, with 30 imprisoned yearly between 2000 and 2014. Homes, livelihoods and relationships were wrecked. Some of those affected took their own lives. Fujitsu recently admitted knowledge of system bugs at delivery in 1999. Notably, the legal framework in England and Wales presumes computer systems are error-free, complicating challenges to computer output. This has serious implications for AI-enhanced systems, as companies integrate AI into decision-making. Current laws risk denying justice to innocent individuals when AI systems err, emphasizing the need for global legal reviews to protect against potential injustices.

Read more at: www.nature.com