Our response to Anthropic’s model welfare initiative - conscious AI

Anthropic is launching a new program to study AI ‘model welfare’.

At PRISM, we’re pleased to see Anthropic leading the way among major AI labs in taking the possibility of AI consciousness seriously. Anthropic’s new 'Model Welfare' research represents a critical step toward understanding and addressing the potential emergence of machine consciousness.

We agree fully with Kyle Fish's assessment that this is an emerging and complex field, demanding significantly more research, interdisciplinary dialogue, and ethical reflection. This effort isn't about asserting that current AI models are conscious; rather, it is about proactively developing the tools, ethical frameworks, and public understanding necessary to responsibly manage such possibilities if and when they arise.

At PRISM, we advocate for a coordinated and collaborative approach that brings together academics, policymakers, civil society, and the broader public. Diverse perspectives and transparent frameworks will be essential in ensuring responsible research and addressing the profound ethical and societal implications of artificial consciousness.

Watch the full interview with Kyle Fish below.