Join RAi UK's upcoming webinar, “In Conversation with: Professor Maria Liakata” to find out more about the RAi UK funded Keystone Project “AdSolve – Addressing socio-technical limitations of LLMs for medical and social computing”.

Register here

This cutting-edge research explores the critical challenges posed by the rapid adoption of Large Language Models (LLMs) particularly in the context of medical and legal use cases.

Despite known limitations – such as biases, privacy leaks, and poor reasoning – LLMs are being deployed without fully understanding their potential consequences. While in some sectors, these risks may be inconvenient, in safety-critical domains like healthcare and legal systems, the impact could be life-altering. For instance, LLMs are already being used by UK judges to summarise court cases, which could lead to serious misjudgements if the AI models reinforce existing biases or misinterpret crucial details.

This project aims to address the socio-technical limitations of LLMs, focusing on two key goals:

1. Developing a robust evaluation benchmark to assess LLM limitations in real-world scenarios, helping regulators and industry partners implement responsible practices.
2. Designing innovative solutions informed by law, ethics, and healthcare experts to mitigate the risks posed by LLMs, ensuring their safe integration into products and services.

Don’t miss this chance to engage live with Professor Liakata, ask your questions, and share your thoughts on the risks posed by LLMs.

The webinar will be chaired by Professor Jack Stilgoe, Deputy Chair Ecosystem Co-ordination Group – RAi UK.

Register here to secure your place

For more information on this project, visit RAi Keystone Projects.