Kavita Abraham
← Back to Encounter
Futures

the commons

A technological intervention for depolarised futures. Common Ground re-engineers the digital architecture of political discourse — replacing outrage-optimised feeds with systems structurally rewarded for finding what people actually share. It doesn't try to change anyone's mind. It changes what the architecture rewards.

Disagreement, without the divide.Common Ground

Depolarised Futures
Concept — Common Ground, a bridging technology
Problem
built for

the extremes

Most social platforms optimise a single metric — engagement — and outrage is reliably more engaging than nuance. The result is a bimodal public square: two loud poles, and an exhausted, disengaged middle. This isn't a flaw in the system. It's the system working as designed.

Why the architecture, not the argument
  1. Bimodal by designEngagement-driven ranking pulls opinion distributions toward two poles rather than a shared centre.
  2. Profit-linkedAttention-data business models make polarising content more valuable to the platform than consensus.
  3. Self-reinforcingEcho chambers narrow the range of views a person even sees, hardening identity around opposition.
Proposed intervention
three moves

not persuasion

Three structural changes, not a persuasion campaign — shifting what's rewarded rather than asking anyone to argue better.

The method
  1. The Bridging FeedModelled on the open-source Pol.is method: surfaces statements gaining agreement across opposing clusters. Reach is earned by cross-partisan validation, not outrage.
  2. The Reframing AssistantA real-time language model rewrites hostile text to keep the point but drop the contempt — naming the legitimate concern underneath before it escalates.
  3. The Assembly EngineAI clusters open-ended arguments into core positions; quadratic voting weights what matters most — without letting majorities silence minority concerns.
Proposed method
how it

works

From disagreement to shared ground
  1. Contribute a viewParticipants submit short, plain-language statements on a live issue — no debate threads to perform for.
  2. Others react, not replyPeople agree, disagree, or pass on each other's statements. No arguing back — this keeps the data honest.
  3. The map appearsStatements cluster by who agrees with whom, revealing the real shape of opinion — including shared ground.
  4. Consensus is surfaced, not manufacturedStatements with cross-cluster agreement rise to the top. Disagreement stays visible — it's just no longer the only thing amplified.
Evidence — a case example
the mechanism

already scaled

"What began as an open-source deliberation tool after Taiwan's 2014 Sunflower Movement became the longest-running national-scale implementation of bridging-based consensus technology in the world."Case study — Pol.is & vTaiwan

Developed by

Computational Democracy Project

Live since

2012

Conversations hosted

Tens of thousands

Participants worldwide

10M+

Status in Taiwan

National digital infrastructure

Sources — Computational Democracy Project · en.wikipedia.org/wiki/Pol.is · arXiv:2502.05017

Imagining a future where technology confronts our echo chambers, and algorithms work intentionally to structurally reverse "unconscious" bias — where AI drives us to systemically see what we already share.Dr Kavita Abraham