Discover AI Pulse
AI-Driven Dialogues for Israeli/Palestinian Peace
AI Pulse connects international stakeholders, civil society, and the broader Israeli and Palestinian publics—groups with few opportunities for dialogue today—to better understand one another and come up with strategies to bring sustainable peace.
How does it work?
In partnership with the high-tech firm Remesh AI, AI Pulse uses AI-driven technology to facilitate large-scale, structured online conversations. Participants share their perspectives in their own language and vote on statements made by others.
Then, AI algorithms analyze responses in real time to identify shared values, areas of division, and actionable insights. Learn more in this helpful tutorial from Remesh AI founder, Andrew Konya.
Our Activities
Ongoing studies turn public voice into actionable insight
AI Pulse is not a one-time study—it’s an ongoing initiative. Our goal is to understand where the public actually stands rather than relying on dominant narratives.
We continuously conduct dialogues and polling exercises with the public, adapting in real time to unfolding events. Our findings are shared with regional and international stakeholders to inform peacebuilding strategies and policymaking.
- Latest Study: May 2026
Finding a Shared Path Forward: Palestinian Public Opinion on Civil Society, External Partners, and a Regional Framework
Our May 2026 AI Pulse study challenges a core assumption of the standard discourse: the vast majority of Palestinians are not hostile to Israeli civil society — they simply haven’t heard of it. Where Israeli grassroots and human-rights organisations are known, they rate at or above Amnesty International and Human Rights Watch as partners in the fight for Palestinian rights. Drawing on a representative sample of 771 Palestinians across the West Bank and Gaza, the study finds 55% support engagement with Israeli solidarity organisations and 67% see a regional peace framework as at least tolerable — but only if statehood delivers genuine sovereignty and the elements of a decent life.
- What's Next
What conditions would move the Israeli public toward a regional framework?
Next in the series, we turn to the Israeli public. Building on our April 2026 conversation, the upcoming study revisits and validates those findings on the questions that shape Israeli readiness for a regional framework: security guarantees, the conditions under which the public would accept a regional framework and a Palestinian state, and attitudes toward settler violence.
- Public
- General Public Perspectives
- Peacebuilding
- Perspectives from the Peacebuilding Field
AI Pulse connects peacebuilders across borders to surface shared values, visions, and strategies for a just and lasting resolution.
- FAQ
Frequently Asked Questions
Select a topic to explore
What is AI Pulse?
AI Pulse is a research initiative by the Alliance for Middle East Peace (ALLMEP), developed in partnership with Remesh, an AI-powered deliberation platform. It runs large-scale AI-assisted dialogues and online polls with Israeli, Palestinian, and international audiences to track how attitudes towards the Israeli-Palestinian conflict and peace evolve over time. The goal is to bring these voices directly into diplomacy and give policymakers and peacebuilders reliable evidence for strategies and action.
What do you mean by "AI-Assisted Dialogue"?
An AI-assisted dialogue is a structured session where participants answer open-ended questions in their own words and vote on each other's responses. AI processes the results at scale, identifying which idea gets the most support across different demographic groups.
What language is AI Pulse conducted in?
Studies are conducted in participants' native languages — Hebrew for Israelis, Arabic for Palestinians, and English for US and UK audiences. Reports are published in English, and some materials are translated into French, German, and Italian.
Is AI Pulse a poll, a survey, or something else?
It's something different. Think of it as a focus group and a representative poll in one. Like a focus group, it lets people speak in their own words, respond to each other, and reveal the reasoning behind their views. Like a poll, it does this at scale, across representative samples, producing clear and comparable numbers. That mix is what makes it unique: the depth of a conversation with the reach and rigor of a poll.
What is AI Pulse’s methodology?
AI Pulse combines open-ended dialogue with polling, run on the Remesh platform. Participants first answer questions in their own words and languages and vote on one another's responses, so the findings capture where a whole group lands rather than a set of individual answers in isolation. Those open-ended contributions then become the statements tested in a second, validation stage, so the research follows what participants themselves raised rather than only what the team anticipated. Consensus is measured across communities: a statement counts as shared only when it wins real support on every side, so a larger group can't outvote a smaller one — and the method is just as clear about where views remain contested.
How do you identify consensus in the polls?
We use a bridging measure called max-min agreement. For each statement, we look at how much each community agrees with it and take the lowest of those figures, so a statement's score is set by its least-supportive group, not by an overall average. Ranking statements this way surfaces the ones that hold up across every community at once, so a larger group can't carry a statement over a smaller one. We then report as consensus the statements whose lowest per-group agreement clears our threshold. The bar can be set high: in our published process with Israeli and Palestinian peacebuilders, many shared statements had 80% agreement on each side. The bridging approach, max-min agreement, was first applied in the initial phase of AI Pulse, documented in a peer-reviewed paper co-authored with researchers from Harvard, MIT, King's College London, and the University of Notre Dame.
Why search for consensus?
Most polling on the conflict measures how divided people are, and asking someone to pick a side can harden the divide rather than ease it. AI Pulse is built to do the opposite. It looks for the agreement that's already there but goes unnoticed when everyone expects only conflict. Every part of the design serves that aim, so the findings show where movement is possible, not just how far apart the two sides stand.
How is AI integrated into AI Pulse?
Participants answer questions in their own words and vote on each other's responses. Because no one votes on every response, an inference model estimates how the full group would have voted on each one. AI then analyzes the results in their original language. It groups responses into themes and distills them into collective statements that capture the group's shared views. AI is also used to generate translations, but native speakers review the AI's output in Hebrew and Arabic to check accuracy.
AI does not design the research agenda, interpret the findings, or replace human judgment.
How do you decide what to study?
Topics follow ALLMEP's strategic priorities, current events, and diplomatic timing. Questions are often designed to reveal hidden consensus where disagreement is assumed, so the findings can offer a concrete roadmap for diplomatic and grassroots activity. Throughout, we consult a wide range of experts and partners (specialists on the topics we study, policymakers, and grassroots organizations) seeking insights to make their own work more impactful, and invite their feedback on what to ask and how to frame it.
How are AI Pulse questions designed?
Questions take several forms: open-ended prompts, polls, and ranking exercises. These are chosen to fit what each part of a study is trying to surface. One approach we rely on is conditional framing: instead of asking "do you support or oppose this?", we ask "under what conditions would you consider this?", turning a yes/no into the terms that might actually make agreement possible. We also use asymmetrical response scales drawn from peace-polling methodology: rather than a symmetric agree/disagree, options run across degrees of acceptability: Essential, Desirable, Acceptable, Tolerable, Unacceptable. This captures not only what people want, but what they can live with, which is often where the real room for agreement lies.
What research tradition does AI Pulse build on?
Using public opinion to support peace talks isn’t new. In Northern Ireland in the 1990s, Colin Irwin ran peace polls that brought the public into the negotiations behind the Good Friday Agreement. AI Pulse works in that tradition: treating public opinion as an input to talks, not just a measure of them. It shares a further goal with deliberative polling, a method developed by Stanford’s James Fishkin, built on the idea that opinions count for most once people have had a real chance to reflect. The methods differ: deliberative polling gathers the same group for a guided event running several days, with briefings and expert sessions; I Pulse adapts that spirit to a single sitting with a fresh sample, easing people in with everyday questions, then letting them weigh each other’s answers.
Has the AI Pulse methodology been used elsewhere?
AI Pulse runs on Remesh, an AI-assisted platform, and builds on its work with the United Nations in peacebuilding since 2020, across more than 60 dialogues in conflict zones including Yemen, Iraq, Lebanon, and Libya. The AI Pulse initial phase is documented in a peer-reviewed paper co-authored with researchers from Harvard, MIT, King’s College London, and the University of Notre Dame.
Who participates in AI Pulse studies?
AI Pulse engages Israeli and Palestinian publics and civil society peacebuilders from the region. ALLMEP also conducts studies with international audiences, including the US, UK, and Canada.
How are participants recruited?
The Israeli public is recruited through two established panel providers: iPanel and Panel4All. For the Palestinian studies, Remesh works with ISEP to obtain representative samples of the Palestinian population. International participants are recruited through Prolific, a platform directly integrated into Remesh. The civil society and peacebuilders community is recruited directly by staff at ALLMEP and in collaboration with partners in the region to get a diverse and comprehensive pool of participants.
How is participant privacy protected?
Participants join the studies via anonymous links. The research team receives no personal details; only an anonymized identifier is recorded on the platform. Only the panel provider can connect an identifier to an individual, and that information is never shared.
What is the average sample size?
Regional studies with Israeli and Palestinian publics typically run around 500 respondents. International studies, such as those with US and UK audiences, tend to reach around 1,000.
Can I participate in an AI Pulse poll?
We hope so. Most participants are recruited through established panel providers. If you're already registered with one, an invitation may land in your inbox. Civil society organizations interested in structured dialogues can get involved through ALLMEP's network. Sign up for the ALLMEP newsletter to stay informed about upcoming studies.
Who uses AI Pulse findings?
Reports are shared with policymakers, diplomats, and heads of state at key diplomatic moments, including at the UN, the Paris Peace Forum, the US State Department, and the UK Foreign Office. They're also distributed to ALLMEP's 200+ member organizations and made public through reports and social media.
How can I access AI Pulse reports and stay updated?
Reports are published on the ALLMEP website. Findings may be cited with attribution to ALLMEP's AI Pulse initiative.
1725 I St NW Suite 300 Washington, DC 20006
(202) 618-4600 | info@allmep.org
©2026 Alliance for Middle East Peace