Examines how the instrumental framing of AI — treating it as a product rather than a legal actor — functions as a necessary fiction enabling existing liability frameworks to operate. Uses Garcia v. Character Technologies, Inc. (2024) to show how competing instrumentalisations allocate responsibility for harm caused by a chatbot, and argues this framing is pragmatically essential for anchoring accountability in human designers.

Proposes a framework for Canadian AI governance that balances innovation with democratic oversight, meaningful public participation, and evidence-based safety measures. Argues for anticipatory governance that serves public good rather than reactive regulation.

Proposes a two-step framework for evaluating anthropomorphic AI design choices following the tragic Character.AI case. Examines when human-like features serve legitimate purposes versus creating harmful illusions, using ChatGPT's "breathing" feature as an example.

Examines accountability in AI development through recent developments at OpenAI and analysis of the UK's interim International Scientific Report on Advanced AI Safety. Explores the tension between innovation and safety in private AI companies.

Explores whether large language models demonstrate Theory of Mind - the ability to understand others' mental states. Examines research showing LLMs performing at or above human levels on cognitive tests and discusses practical applications in healthcare and decision support.

Examines inadequate whistleblower protections for AI safety concerns following an open letter from OpenAI and Google DeepMind employees. Analyzes gaps in UK whistleblowing law and argues that internal reporting processes may be the most realistic solution.