What — AI ethics tries to achieve
Concepts are power. Shared concepts help us share and build power together.
People or things are aligned insofar as they can be understood to share goals. Individuals, organizations, communities, countries, and AI systems can all be more or less aligned, within themselves and between each other. They can share more or fewer goals, and they can share those goals more or less deeply. All else being equal, sharing goals more deeply increases alignment.
Groups are sets of individuals. Individuals can belong to many groups. The kinds of groups that we focus on here are ones that have a shared sense of what they should do, and how they should be, which is their ethos. Successful human groups consistently depend on a shared ethos to help them create and maintain their identity, and coordinate their actions.
We do things ethically when we explicitly do them in ways that align with the ethos of at least one group. Whenever we talk about what we should do, or how we should be, we are doing ethics. Things are ethical insofar as they align with the ethos of a group, and more broadly to the degree that they respect the ethical agency of all.
The Core of the AGF
Cultivate Human Groups Ethically
The AGF is designed to help you align groups ethically, so that you can ethically align AI systems to yourselves.