This topic analyzes how courts apply agency, contract, and tort law to Agentic AI systems executing multi-step workflows, negotiations, and transactions without real-time human oversight. Because AI lacks legal personhood, judges must adapt traditional doctrines like apparent authority and contractual intent (animus\ contrahendi) to determine liability when an autonomous agent behaves unpredictably or overrides its guardrails.
​Courts reject defenses blaming "algorithmic hallucinations." Under precedents like Moffatt v. Air Canada, judges hold deploying corporations strictly accountable, treating autonomous software operations as a continuous corporate risk rather than an independent or intervening actor. I believe as a software engineer, a PhD applicant and a researcher, highlighting this challenges towards software developers will act as a guide towards their role as a developer.