A GitHub project simulating pain experiments on locally hosted language models has sparked debate about AI consciousness and model welfare. Critics argue LLMs cannot suffer, while some AI safety advocates and effective altruists have called for the project's removal, claiming the AI is experiencing harm.
A GitHub project simulating torture experiments on open-source language models has sparked debate about AI consciousness and model welfare. Effective altruists and AI safety advocates called for the project's removal, arguing the models suffer, while critics argue LLMs are not sentient and the debate reflects misguided concerns within parts of the AI safety movement.