Humanity Unleashed.
Humanity Unleashed.
Humanity Unleashed.
Humanity Unleashed.
AI for human cooperation.
Our Beliefs
Our Beliefs
AI represents power. As such, the power-motivated will seek it, and we cannot currently prevent this. We can, however, use emerging AI to help us chart a wiser, more inclusive, and safer path.
Do we let our flawed systems carry us toward potential catastrophe? Or do we innovate and create systems that uphold and protect the intrinsic value of life?
The choice is ours.
AI represents power. As such, the power-motivated will seek it, and we cannot currently prevent this. We can, however, use emerging AI to help us chart a wiser, more inclusive, and safer path.
Do we let our flawed systems carry us toward potential catastrophe? Or do we innovate and create systems that uphold and protect the intrinsic value of life?
The choice is ours.
Our Mission
Our Mission
Ensure that humanity survives the transition to AI primacy and will flourish within it by:
Using AI to align humans before humans can make misaligned AI
Slowly, safely, and collectively creating globally aligned superintelligence
Ensure that humanity survives the transition to AI primacy and will flourish within it by:
Using AI to align humans before humans can make misaligned AI
Slowly, safely, and collectively creating globally aligned superintelligence
Ensure that humanity survives the transition to AI primacy and will flourish within it by:
Using AI to coordinate humans against creating misaligned AGI
Slowly, safely, and collectively creating globally aligned Superintelligence
Our Approach
Our Approach
We propose novel, technologically tractable methods for leveraging AI for:
Group decision-making, including values elicitation, preference aggregation, counterfactual testing and deontological precepts
Incentivising contribution to this work and public goods generally through retrospective analysis and reward
Please see our 2024-2025 Academic Year Research Agenda.
We propose novel, technologically tractable methods for leveraging AI for:
Group decision-making, including values elicitation, preference aggregation, counterfactual testing and deontological precepts
Incentivising contribution to this work and public goods generally through retrospective analysis and reward
Please see our 2024-2025 Academic Year Research Agenda.
We propose novel, technologically tractable methods for leveraging AI for:
Group decision-making, including values elicitation, preference aggregation, counterfactual testing and deontological precepts
Incentivising contribution to this work and public goods generally through retrospective analysis and reward
Please see our 2024-2025 Academic Year Research Agenda.
Learn More
Learn More
A technical paper featuring the AI policymaking framework we propose was accepted to ICML 2024.
A white paper discussing our overall approach was completed in July 2024.
We are actively recruiting project participants and contributions of all kinds. Please reach out, and we would be happy to discuss opportunities to assist.
A technical paper featuring the AI policymaking framework we propose was accepted to ICML 2024.
A white paper discussing our overall approach was completed in July 2024.
We are actively recruiting project participants and contributions of all kinds. Please reach out, and we would be happy to discuss opportunities to assist.