July 2024

AI has been unleashed on the world, and it is only a matter of time before it surges from a valuable tool to a new paradigm of intelligence vastly exceeding our own. We must now unite and unleash the collective moral and technical capacities of humanity toward safely developing and integrating AI into our society. If we fail in this, we will instead see these tools used to amplify our worst qualities as we struggle for outright dominance.

Artificial Intelligence doesn’t yet inherently seek power, but humans do. To ensure that Superintelligent AI is aligned to our collective interests, we must first ensure that the humans building it are. Existing economic and governance systems are not up to this task and are already pushing AI development in hazardous directions.

Fortunately, there is a solution: leveraging pre-superintelligence AI to help align human interests can allow us to work together as a global community and take the time needed to get this right. Humanity Unleashed, a 501(c)(3) nonprofit organization, is engaged in constructing such an alignment system that can unite humanity around the intrinsic value of life and reward contribution toward public goods.

Our approach consists of four interrelated development goals, all of which leverage emerging AI capacities to extend the realm of the possible: 1) elicit foundational values from individuals through new methods of voting and interactive systems, 2) develop automated policymaking capacities that can efficiently and equitably find tractable paths forward from discordant values, 3) value contributions to public goods, inclusive of the work proposed here, by enabling retrospective analysis, and 4) ensure demonstrable positive impact from our research and our organization generally throughout the process.

This development work is building toward an integrated system that will remain durably committed to the global welfare, advancing our collective interest and rewarding those who assist. This paper lays out the context and rationale for taking action and our plan for accomplishing these goals.

The march of technological progress has finally delivered a world in which it is not possible to defer the inherent dysfunction of our collaborative systems any longer. There are no more frontiers. There are no escape hatches or fall back positions.

We are taking a final, principled stand, and we hope you will join us.