Why it matters
Arjun Krishna and collaborators measure fictional dependency generation across eleven models and Python, JavaScript, and Rust tasks. They find that package-hallucination behavior varies with the model, language, size, and request specificity, creating a supply-chain opening when an attacker registers a plausible package name suggested by an AI coding system.
My takeaway: Treat every AI-suggested dependency as untrusted until the registry owner, publication history, source repository, license, integrity hash, and malware checks are verified. Route installs through an approved registry, lock and hash dependencies, block first-seen packages by default, and regression-test coding models across every supported language and task family.