Non-Product Defense Case Library

Software, AI, and the Boundaries of Strict Liability

An Interactive Reference for Product Liability Defense Counsel

TRUE GOLD Evidence 50 Sources 3,848 Warehouse Items 32 Cases

Filter Cases

Search
Posture
Category
Analytical Frameworks From Warehouse Sources

Sharkey Thesis — Products Liability for AI

SOURCE: lawfaremedia_org_7d9e113add (158 pins)
1.Historical Arc: Products liability has always evolved with transformative technology. Automobile → privity eliminated. Mass production → strict liability. Complex products → design defect doctrine. AI is the next chapter.
2.Abandon Tangibility Test: The products/services distinction and tangibility metric are ill-suited to AI. Courts should ask whether products liability is justified by functional policy: mass production, widespread harm risk, deterrence, loss spreading.
3.Mass-Market LLMs as Products: ChatGPT, Claude, Gemini resemble off-the-shelf products — identical source code mass-distributed. Fine-tuned agents may resemble services. The mass-market/custom distinction is key.
4.Cheapest Cost Avoider: Courts should impose liability on the party best positioned to avoid or mitigate harm, bypassing formalistic product/service distinctions entirely.
5.Postsale Duties: AI developers cannot red-team predeployment and wash their hands of post-deployment risks. Products liability imposes continuing duties to monitor and warn.
6.Component Liability: Foundation model developers may be liable for downstream harms from modified versions — analogous to Air & Liquid Systems v. DeVries "bare metal" doctrine.
7.Information-Forcing: Products liability forces risk disclosure for opaque AI systems during the current "transition period" before ex ante regulation.

Jones Walker — Vendor Liability Squeeze

SOURCE: joneswalker_com_b26dca604b (49 pins)
Courts expand vendor accountability (Mobley agency theory, strict liability exploration) while contracts shift liability to customers (88% of AI vendors impose liability caps; only 17% provide regulatory compliance warranties). Deployers face a "liability squeeze" — responsible for AI failures they cannot audit, control, or understand.

AI LEAD Act — Statutory Framework

SOURCE: library_law_uic_edu_cfa533f68c (102 pins)
S. 2937, 119th Congress (2025). Bipartisan (Durbin/Hawley). Would classify AI systems as "products" by statute, creating federal cause of action for design defects, failure to warn, breach of warranty. Extends liability to deployers who substantially modify AI. Prohibits unreasonable liability waivers. If enacted, eliminates the non-product defense entirely for AI systems.