Software that mimics human society is being tested to see if it can help prevent religious violence.
Researchers used artificial intelligence algorithms to simulate actions driven by sectarian divisions.
Their model contains thousands of agents representing different ethnicities, races and religions.
Norway and Slovakia are trialling the tech to tackle tensions that can arise when Muslim immigrants settle in historically Christian countries.
The Oxford University researchers hope their system can be used to help governments respond to incidents, such as the recent London terror attacks.
However, one independent expert said that the tool needed more work before it could be used in real-life situations.
“This could be an extremely useful research project when it reaches maturity as a thought tool for analysing factors involved in religious conflict,” said Prof Noel Sharkey.
The research, published in the Journal for Artificial Societies and Social Stimulation, indicates people are a peaceful species by nature.
Even in times of crisis, such as natural disasters, the simulated humans came together peacefully.
But in some situations the program indicated people were willing to endorse violence.
Examples included occasions when other groups of people challenged the core beliefs that defined their identity.
The research team drew on the Northern Ireland Troubles to programme conditions that saw what the researchers call “xenophobic social anxiety” escalate into extreme physical violence.
The conflict – which involved political and cultural factors in addition to religion – spanned three decades and claimed the lives of approximately 3,500 people.
The computer model also used scenarios based on the 2002 Gujarat riots in India. Two thousand people died during three days of inter-communal violence between Hindus and Muslims.
Research author Justin Lane said: “To use AI to study religion or culture, we have to look at modelling human psychology because our psychology is the foundation for religion and culture.
“The root causes of things like religious violence rest in how our minds process the information that our world presents.”
The results suggest the risk of religious conflict escalates when a group’s core beliefs or sacred values are challenged so frequently that they overwhelm people’s ability to deal with them. But even then, anxiety only spills over into violence in about 20% of the scenarios modelled.
“Religious violence is not our default behaviour – in fact it is pretty rare in our history,” said Mr Lane.
“It is only when people’s core belief systems are challenged, or they feel that their commitment to their own beliefs is questioned that anxiety and agitations occur.
“We might be able to trick our psychology into accepting others as part of our group when we’d otherwise be triggered toward more primal fears,” he added.
The researchers believe one answer to reducing the risk of religious violence and terrorism is to create situations that stop people seeing outsiders as a threat.
The most risky situations are when the difference in the size of two different religious groups is similar and people encounter “out-group members” more regularly, perceiving them as dangerous.
The encounters need not be face-to-face. It could be that the threat is brought to someone’s attention through conventional and social media.
“We appear to live somewhat in our own information bubbles, but we still receive a lot of information about out-group members and that information appears to trigger our psychology even if there isn’t actually a real person there,” Mr Lane warned.
“Just the idea of a threat can be as powerful as a real threat to elicit a reaction.”
Can artificial intelligence help stop religious violence?