Researchers from Google DeepMind, in collaboration with other experts, have developed an AI system called the 'Habermas Machine' to facilitate democratic deliberation. The system, which is a large language model (LLM), has been shown to help find common ground on politically divisive issues in the UK better than human mediators, according to a new paper published in Science Magazine. The research involved a series of experiments with over 5000 participants, demonstrating the efficacy of AI-mediated deliberation.
One of my favourite papers in recent months. 'Positive-sum symbiosis' in action. The authors conduct a series of experiments with over 5000 participants to evaluate the efficacy of AI-mediated deliberation. The LLM-powered ℌ𝔞𝔟𝔢𝔯𝔪𝔞𝔰 𝔐𝔞𝔠𝔥𝔦𝔫𝔢 uses supervised… https://t.co/sfZm9t6D5Q
DeepMind researchers find LLMs can serve as effective mediators @sciencemagazine https://t.co/JZ8BfxLWzL
Habermas Machine (an LLM) helps find common ground on politically divisive issues in the UK better than human mediators https://t.co/PwYWRnapbu @ScienceMagazine https://t.co/XnaccYrX9e