Making democracy work isn’t easy, as recent events have made clear. Some critics would argue that technology is making it worse. But one startup is hoping that AI could help bridge some differences instead of widen them.
“I had an a-ha moment one day when I realized people are asking AI to explain something like they’re five years old,” Tomy Lorsch, co-founder and CEO of ComplexChaos, told TechCrunch. “What if we use it as a facilitator to help people understand each other and find common ground?”
He and co-founder Maya Ben Dror are developing tools to help people arrive at a consensus. One of their first test cases involved climate negotiations, but it really doesn’t matter what the issue is. Their goal is to foster cooperation and shorten the time it takes for groups to come to agreement.
“Everyone is building software for collaboration like Slack, Google Docs, whatever,” Lorsch said. “Cooperation is a different piece.”
Facilitating cooperation isn’t something that scales well, he said. Typically, trained facilitators will spend time with groups to help them arrive at a consensus, but that process can slow down when negotiations or preparations happen across time zones or even in different rooms.
Lorsch was buoyed by a recent LLM developed by Google called the Habermas Machine, which was developed explicitly with that goal in mind. “This is basically an AI that generates group consensus statements where people feel represented both by majority and minority point of view,” he said.
Lorsch and Ben Dror recently trialed their startup’s tool to help young delegates from nine African nations prepare for climate-related negotiations at a United Nations campus in Bonn, Germany. The tool incorporates both Google’s Habermas Machine and OpenAI’s ChatGPT to generate questions, come up with goals for conversations, and help summarize long documents.
The goal, Ben Dror said, is to help the delegates arrive at consensus as a bloc before they begin negotiations with others.
Ideally, the tool would help speed things during negotiations, too. When blocs, or delegates from groups of aligned countries, encounter new information in the process of a large negotiating session, they often need to regroup to process the new information. “Blocs are usually the reason why negotiations have to stop. The bloc has to come out, renegotiate, reposition, and then go back in. And that creates a lot of friction,” Ben Dror said. ComplexChaos hopes that its tool can help shorten that time.
In the trial with the delegates from African countries, ComplexChaos said that participants reported up to a 60% reduction in the time it took to coordinate, and that 91% of participants said that the AI tool helped them see perspectives they would have otherwise missed.
ComplexChaos is also pitching its cooperation tool to companies, including tech companies and large consultancies. “Strategic planning by AI is basically the same problem,” Lorsch said. “The annual strategic planning process of most companies takes about three months of the year with back and forth negotiations, multi-layer, across time zones, across teams, and so on and so forth.”
But Lorsch and Ben Dror are most enthusiastic when talking about climate negotiations.
“If AI can shorten these processes, simplify them, then we’d be so much better off. Not just for climate, for anything sustainability, for any big challenge that we’re facing,” Ben Dror said.
