UN using AI to help resolve Israel-Palestine conflict

The United Nations is deploying AI models in an effort to reach a peaceful solution to the Israeli-Palestinian conflict. But given the recent horrific escalation in hostilities, likely few will be holding their breath.

The United Nations Development Program (UNDP) has contracted Slovakia-based AI startup CulturePulse to “bolster its peacekeeping client roster” and “use AI to understand critical issues in the Israel-Palestine conflict,” according to a statement released by the company this week.

CulturePulse and UNDP have agreed to collaborate on a five-month trial project to develop “computational models of causal dynamics in the Palestine-Israel socio-ecological system” to help the UN achieve its goal of promoting stability in the region.

Launched in August, the project was announced on October 10th in the wake of Islamist militant group Hamas’s attack on the Jewish state, the Palestine-Israel Virtual Outlook Tool (PIVOT) will test out potential interventions through digital simulations before “trying them out in the real world.”

It will be led by CulturePulse CEO Dr Justin Lane and Professor LeRon Shults of the Institute for Global Development and Social Planning. The company is claiming it to be the first of its kind with the United Nations.

High hopes for digitally driven peace

Despite the horrendous casualties on either side, with terror group Hamas and the Israeli Defense Force each killing hundreds of people including children in the days since the former’s surprise offensive on Saturday, Dr Lane told Cybernews that he is optimistic that the AI-led project can help to bring about conflict resolution in the long term.

“We found that there is a psychological signature that is really common in extremists called ‘identity fusion’ that was often the result of being a victim of terrorism as well,” he said, when asked about the mindset that informs terrorist organizations.

“Psychological studies show that extremist actions can create extremism and begin to explain the cycle of violence,” he added. “AI modeling helps us to uncover the findings about the issue of identity fusion and extremism.”

Dr Lane hopes that CulturePulse’s algorithms, when combined with studies of political and religious extremists and psychological profiles of paramilitary and military organizations, will shed greater light on seemingly intractable problems such as the Israel-Palestine conflict by mapping the mindsets of antagonists.

“Our system is uniquely capable of figuring out what the beliefs an individual holds in comparison to their social group,” he said. “Where their personal sort of identity, like what it is that makes them tick, is something that they share socially.”

He added: “It's not just that they have a belief and go off and do something extreme about it. That is one form of radicalization that we can pick up, but this other form is particularly strong. This identity fusion is when you're so tied to a social group that you're willing to fight and die for them.”

Can AI unpick human conflicts?

Dr Lane claims that he and LeRon have developed AI conflict tracking software that “understands these core issues of social and personal identities” that will help to “craft messages that can strategically address issues of extremism.”

CulturePulse previously worked with Cambridge University to develop AI models that simulate and analyze social conditions surrounding conflict and peace in Northern Ireland, which was riven by sectarian violence for more than three decades resulting in thousands of deaths.

This involved distilling millions of articles from a vast database of human behavior broken down into 80 sub-categories of culture, psychology, and morality underpinning the conflict.

CulturePulse repeated this approach to seeking resolutions to the conflict in the Balkans, in particular Bosnia and Herzegovina, and addressing the Syrian refugee crisis, with a focus on integrating these into host countries “by understanding moral differences and values.”

“In some cases, you want to de-radicalize someone and sometimes you need to do that slowly, right?” he added. “Sometimes that's a long process where you need to work around these key trigger terms and concepts that are going to set somebody off. And in a debate, you cannot really activate them, you need to kind of talk to them in a calm way.”

“Our system, because it maps where those triggers are, also provides the map to get around it so that you can connect the different ideas like, you know, peace and economics and different aspects of negotiation, and allows you to get around those key triggering ideas.”

NB: This article was amended on October 16th to stress the difference between machine learning and artificial intelligence, and also to correct a factual error: the PIVOT project was in fact first tabled in August, and not after the Hamas attacks on Israel, as was previously reported.