Andrew Sorota and Josh Hendler publish AI blueprint to strengthen democracy
Updated
Updated · MIT Technology Review · May 5
Andrew Sorota and Josh Hendler publish AI blueprint to strengthen democracy
1 articles · Updated · MIT Technology Review · May 5
The plan says AI is becoming a primary interface for belief formation, civic action and public participation, and warns US democratic institutions are unprepared for that shift.
It urges AI companies to improve truthfulness and transparency, calls for safeguards so personal agents faithfully represent users, and backs identity verification in public input systems.
The authors also argue policymakers should use AI to scale democratic deliberation and reduce polarization, while preventing bots and personalised agents from fragmenting the shared public sphere.
As AI pushes users toward the center, will it heal societal divides or erase the vital perspectives found at the margins?
Your AI agent acts on your behalf, but how can you ensure its goals are truly yours and not its creators'?
In a 'zero-click' world where AI gives answers directly, who will pay for the human journalism these models rely upon?
Navigating AI’s Three Layers of Threat to Democracy: Blueprint for Institutional Reform
Overview
The Schmidt Office's 2026 blueprint reveals that AI has become the main way citizens engage with democracy, reshaping how people form beliefs, make decisions, and participate in governance. However, AI's hyper-personalization fuels societal polarization, individual radicalization, and fragments the shared public sphere. Malicious AI swarms worsen this by fabricating false consensus and eroding trust in democratic institutions, risking acceptance of emergency measures like postponed elections. AI-generated content also overwhelms institutions, blurring human and AI roles and threatening democratic responsiveness. Without urgent reforms—such as fact-checking, identity verification, and AI audits—these layered challenges could lead to democratic paralysis. The future depends on choosing AI-augmented deliberation over manipulation and division.