Using Consensus Mechanisms as an approach to Alignment — LessWrong

 Part 1 - Introduction Mechanism Design - Programming human behavior through carefully designed incentives. Humans are goal-oriented machines that react to incentives. Incentives are produced by economic games that can be crafted and engineered. So we humans do something to reach a goal, and one of the most unifying goals is money. So we desi

... See more

The tipoff to

the nature of the AI societal risk claim is its own term, “AI alignment”. Alignment with what? Human values. Whose human values? Ah, that’s where things get tricky.

AI alignment is just the problem of AI bias writ (terrifyingly) large: We are assigning more societal decision-making power to systems that we don’t fully understand. D o these two groups of experts charged with making AI safe actually get along? T hese are two camps, and they’re two camps that sometimes stridently dislike each other.

AI that can start to be an AI scientist and self-improvement could help us solve the hard alignment problems that we don’t know how to solve. The alignment problem is: how do we build AGI that does what is in the best interest of humanity? How do we make sure humanity gets to determine the future of humanity. It is also interesting to note how this... See more

The field of game theory models interactions between rational decision makers, identifying dilemmas of cooperative, zero-sum, and symmetric games and their inverse. As a counterforce, decentralized autonomous organizations broadly apply mechanism design through economic incentives or social norms to achieve rough consensus and running orgs. Headles... See more

The intersection of AI and DAOs - Agent-based modeling is an area of ai research - Similar to governance mapping (which approaches this problem from the realm of economics). - Look at all of the agents in a system and what their capacities and goals are - Use massive simulations to understand the ways that participants could act - Or teach them how... See more