Reverse Alignment Coalition

= " reverse alignment: the deliberate redesign of institutions, norms, skills, and governance so that societies can absorb AI safely and share its gains widely. " [1]

URL = https://reversealignment.ai/

Context

Georg Diez and Felix Beer:

"AI accelerates the need to rethink the foundations of democratic life. Information is easier to produce and analyze. Increasingly complex tasks can be delegated to machines. Change itself accelerates. AI is entering governments, firms and everyday life, it is reshaping how knowledge is produced, how people coordinate and how authority is exercised.

These shifts create new pressures on democratic institutions, from election integrity to agentic flooding of public services and regulatory manipulation. They also make possible forms of collective action that were once hard to sustain, or even to imagine. A recent example: In one study, an AI mediator called the Habermas Machine helped divided groups find common ground.

Regulating AI and adopting it responsibly within existing institutions are necessary responses. But it is not enough. AI changes the conditions on which those institutions depend. If governments gain greater capacity to act while public oversight remains weak, technological progress can concentrate power. Meeting this challenge requires redesigning institutional infrastructure alongside technological advances.

The Reverse Alignment coalition names and maps the scale of this “socio-technical grand challenge” for institutional invention. Its agenda ranges from establishing identity and protecting privacy to supporting collective judgment, repairing trust, and rebuilding checks and balances for an age of intelligence explosion. Meeting these challenges requires large-scale investment and coordinated action across government, civil society, academia, and frontier labs.

The result will be a variety of new operating systems for democracy.

We understand democracy as collective self-government among political equals. This requires both public agency and institutional capacity as people need meaningful opportunities to form judgments, organize, influence decisions and challenge power; and institutions need the capacity to act on collectively authorized choices.

The new setup will be a different operating system from the one we have; the ultimate democratic test is whether people gain greater ability to shape their shared life.

We propose to build towards such an operating system, recognizing that what Audrey Tang calls MaxOS is already in a lot of ways in place. The question is whether technological change strengthens or undermines this democratic agency. Progress is not on autopilot: greater technological and administrative capacity does not automatically strengthen self-government. That makes democracy’s institutional infrastructure a subject of deliberate design."

(https://informationaldemocracy.substack.com/p/operating-systems-of-democracy?)