Traditional principles of misrepresentation depend on human intention & agency. The rise of autonomous AI systems challenges both concepts in fundamental ways, writes Mary Young
- As artificial intelligence (AI) chatbots and large language models increasingly provide information and advice, English law is being forced to consider who may be legally responsible when those outputs are false or misleading.
In January 2026 the UK Jurisdiction Taskforce launched a consultation on a legal statement it had prepared on liabilities for artificial intelligence (AI) harms, considering what rights of recourse someone harmed by AI might have, but also against whom those rights might be exercised.
This article considers misrepresentation, and how the laws of England and Wales might approach representations made by automatons such as large language models (LLMs) (for example ChatGPT) or AI chatbots.
For there to be an actionable misrepresentation, there must be a false statement made by party A (or their agent) which induces party B to enter into a contract. While there is a degree of anthropomorphism in relation to LLMs (they often refer to themselves in the first person
