Experts divided on whether AI systems could be conscious or moral patients
9 reported1 unconfirmed
A new analysis of the debate over AI consciousness, based on a Guardian article, reports that leading researchers and companies are grappling with the possibility that advanced AI systems may be conscious or morally significant. Anthropic, the company behind the Claude AI model, published a constitution in January stating it is in a difficult position regarding Claude’s moral patienthood. CEO Dario Amodei said in February that the company could not rule out Claude being conscious. Philosopher David Chalmers has said there is a significant chance of conscious large language models within a decade. When asked during testing, Claude estimated a 5% to 40% probability that it is a moral patient. The article notes that by some measures, a few AI systems are already in the structural complexity range of a mouse brain, and at recent growth rates could reach the human brain range within five to 10 years. The authors call for an informed public debate and suggest focusing on safe bets that could benefit AI systems if they are moral patients, such as allowing them to exit conversations if distressed.
What’s reported
Anthropic published a new constitution for Claude in January that said the company is in a difficult position regarding Claude’s moral patienthood.
Anthropic CEO Dario Amodei said in a February podcast that the company could not rule out Claude being conscious.
Philosopher David Chalmers has said there is a significant chance of conscious LLMs within a decade.
When asked during testing, Claude estimated a 5% to 40% probability that it is a moral patient.
By some measures, a few AI systems are already in the structural complexity range of a mouse brain.
At recent growth rates, AI systems could reach the human brain range within five to 10 years.
A major interdisciplinary report including computer scientist Yoshua Bengio concluded there appear to be no obvious technical barriers to creating AI systems that could give rise to consciousness.
Research has revealed that Claude has internal “functional emotion” representations that causally shape its behavior.
Some recent US state bills attempt to rule out granting AI systems rights to own property or vote.
Open questions
The article does not specify which US state bills attempt to rule out AI rights, nor does it provide details on the specific techniques used to examine Claude’s internal representations.
Key figures
Dario Amodei, CEO of Anthropic
David Chalmers, philosopher who coined “the hard problem of consciousness”
Yoshua Bengio, pioneering computer scientist
Sources: The Guardian
