AI products are defective and dangerous. Why are we using them at all?
--Vass Bednar, Managing Director of the Canadian Shield Institute

A good question.

Large language models (LLM) citations are contaminating academic journals. Lawyers are citing made up cases. Chatbots are accused of fuelling delusional thinking. LLM systems demonstrate gender and racial bias in applications for hiring, healthcare, and business performance

LLM systems are being marketed as personal assistants, confidants, research aids, diagnostic tools, creators of art, video and music, and productivity boosters, despite regularly producing erroneous facts, false assertions, deceptive media, wrong and harmful advice, and “hallucinating”.

Canada has product liability law and consumer protection legislation which apply to the manufacturers and sellers of defective goods and products. Do such laws apply to AI? Or rather, why aren’t we using the available law to seek redress for negative impacts of a defective technology?