Experts with international experiences in reducing election violence share their lessons for the United States.

Technology can be a transformative tool in addressing electoral violence, said a panel of violence prevention experts. They highlighted some promising examples of effective interventions to reduce violence ahead of electoral processes but also pointed out that novel technologies like AI might present challenges.

The discussion took place on November 2 at George Washington University’s Elliott School of International Affairs. It was hosted by the Institute for Public Diplomacy and Global Communication and moderated by Peace News director Babak Bahador.

Susan Benesch, founder of the Dangerous Speech Project, and Theo Dolan, Digital Technology and Civic Activism Advisor at USAID, pointed to Kenya as an interesting example of how media can support peaceful elections. The East African country faced intense violence after the 2007 presidential election. Incumbent President Mwai Kibaki was reelected but faced allegations of electoral manipulation from opposition figures and international observers. 

The election campaign was defined by polarization and hate speech between Kenya’s various ethnic groups. Bensesch said that the election’s rhetoric transformed the way Kenyans understood their own identity: “This campaign was so successful that you could ask people “Who are you?” and instead of saying “I am Kenyan” they would name their subgroup, their particular ethnic group.” Over 1,000 people died due to the post-election violence and almost 700,000 were displaced.

With tensions high before the 2013 election, peacebuilding practitioners were motivated to experiment with different methods. Benesch cooperated with Vioka Mahakamani, a popular Kenyan courtroom comedy drama, to produce four episodes warning about the dangers of violent speech. Researchers at the University of Pennsylvania later found that the episodes were effective in raising awareness about the dangers and legal consequences of incitement to violence.

Kenya has become a forerunner in using technology, especially text messaging, to prevent violence, said Dolan. He explained how, following the lead of local organizations like Una Hakika and Sisi Ni Amani, the United States Institute of Peace’s PeaceTech Lab deployed texting as a tool for conflict prevention during the 2017 election. These projects all allowed people to send texts about instances of violence or misinformation to a central platform, where fake stories could be debunked and researchers could compile detailed information about political violence in the country. Dolan emphasized that local community leaders and networks remained crucial in preventing electoral violence.

The traditional division between international and domestic electoral violence prevention work is artificial, said Vasu Mohan, Senior Country Director for Nepal and India and Senior Advisor for Conflict and Displacement at the International Foundation for Electoral Systems. He argued that the U.S. had a lot to learn from efforts to prevent election violence abroad. 

Mohan also emphasized the importance of working with local thought leaders to prevent election violence. He pointed to the success of peacebuilding NGOs in the 2018 gubernatorial election in the Indonesian province of West Kalimantan. The region experienced intense political polarization between its Muslim Malay and Indigenous, mostly Christian communities, but civil society organizations worked directly with religious leaders to train them to use counterspeech to dampen the effectiveness of hate speech and disinformation. 

Mohan said that the region also benefitted from fact-checking from Mafindo, a grassroots NGO carrying out fact-checking in Indonesia. The country also had an independent electoral authority that effectively cooperated with civil society. This whole-of-society approach resulted in very little political violence despite widespread incitement to violence, and became a model for future elections in Indonesia.

The panel took several questions on a variety of topics. When asked how artificial intelligence (AI) and similar technologies could be incorporated into violence prevention, Benesch said that more research was necessary to determine if AI-generated counterspeech would be helpful or detrimental. She proposed that bots could be used to produce mass counterspeech: “It’s possible that people will in fact be more responsive to counterspeech produced by AI and more willing to engage with it than counterspeech produced by humans, in part because you can’t infuriate a bot.” 

Benesch admitted that this could be counterproductive, saying that people are more inclined to believe known and trusted sources and leaders.

Dolan agreed with this approach and said that the tools of disinformation could be turned against it: “We tend to be quite slow in learning from malign actors, and we can use some of the tactics that they employ also to our benefit.” However, he argued that AI large learning models (LLM) need to be trained on more diverse data, as most of their information is currently from the global north.

Media literacy remains a powerful preventive tool for violence prevention, according to Mohan and Dolan. Mohan argued for a model of civic education that integrates international human rights and conflict resolution training at every level of education.

“It’s not just media literacy in the sense of just maybe to be able to detect AI-generated content or to detect misinformation, disinformation,” said Mohan, “but also I think parallel to that there has to be something about ethics, about media ethics, and the idea of social cohesion.”

When asked about how to hold political leaders who make non-violent pledges accountable, Benesch said that countries with more powerful parties had an advantage. If parties have an incentive to stick to non-violent language, they can control their candidate’s messaging and ensure they follow up on their pledges. Benesch and Mohan both also said that institutions that can enforce rules or shame candidates, such as independent electoral commissions and competent and professional security forces, are critical.

The panelists argued that in low-technology environments, art can be an effective tool for violence prevention. Mohan said community-based theater was a promising intervention, and Benesch pointed to graffiti artists creating murals against political violence. Dolan added that radio, still the main mass media technology in many global south countries, remained effective in some contexts.

The panelists were also asked about the role of age and gender in violence prevention. Benesch pointed out that young men commit the majority of political violence, especially when their social circle seems to condone that violence. Fear, she argued, is a central driver of political violence, and peacebuilders should address messaging that preys on people’s fear of public servants such as election workers and emergency services. In the U.S., conspiracy theories targeting election workers have become widespread. False rumors about the Federal Emergency Management Agency (FEMA) have also emerged after the devastating hurricanes that affected the southern U.S. in October and September.

Mohan said that young leaders and influencers are key to reaching out to young men, but that women leaders should also be integrated into violence prevention efforts. Too often, women, minorities, and people with disabilities are sidelined during anti-violence programming, he said.

Technology can have regressive effects and facilitate gender-based violence, argued Dolan. He said that when women in public spaces are targeted, this can have a chilling effect on their speech.