Prof. Eliza MIK presented at the UNU Macau AI Conference 2024, which took place in Macau SAR on 25 April 2024 and hosted over 120 speakers from leading tech companies, academia, and civil society.

Prof. Mik’s presentation, titled “Dangerous Liaisons: Large Language Models and Access to Justice”, addressed the risks and benefits of deploying large language models in situations where users may lack the competence to evaluate the quality of the generated output and to query the model with a correctly formulated legal question. In a panel on AI trustworthiness and regulation, Prof. Mik emphasized that when evaluating the reliability of LLMs, it must be acknowledged that the detection or very existence of hallucinations (i.e. incorrect output) may be a function of the user’s competence.