A new computational model tries to find the overlap between two populations who mostly agree on nothing else: what peace terms Israelis and Palestinians might both accept.
Researchers introduce a quantitative bipolar argumentation framework that models how each side reasons about specific clauses in a peace agreement, not just which clauses appear. Merging the two frameworks lets the system compute what the paper calls a Zone of Possible Agreement, a set of terms both populations could plausibly accept. The team tested the approach on the Israeli-Palestinian conflict, drawing on existing survey data plus additional data retrieved with help from a large language model. The paper frames this as preliminary work, with the Zone of Possible Agreement identified through theoretical analysis and early experiments rather than a finished negotiating tool.
Most modeling of peace processes treats public opinion as a fixed poll number, support or opposition to a given clause, and misses the reasoning behind it, which is exactly where negotiators tend to get stuck. By formalizing citizens' arguments for and against specific provisions, this approach could help conflict-resolution teams spot compromise language before it fails in public rather than after.
The catch is the data: a model like this is only as good as the survey responses and LLM-retrieved opinions feeding it, and reducing a decades-long conflict's contested narratives to argumentation graphs is the kind of computational neatness that real-world talks tend to resist.