Participants were divided up into six-person groups, with one participant in each randomly assigned to write down statements on behalf of the group. This person was designated the “mediator.” In each round of deliberation, participants were presented with one statement from the human mediator and one AI-generated statement from the HM and asked which they preferred.
Greater than half (56%) of the time, the participants selected the AI statement. They found these statements to be of upper quality than those produced by the human mediator and tended to endorse them more strongly. After deliberating with the assistance of the AI mediator, the small groups of participants were less divided of their positions on the problems.
Although the research demonstrates that AI systems are good at generating summaries reflecting group opinions, it’s vital to remember that their usefulness has limits, says Joongi Shin, a researcher at Aalto University who studies generative AI.
“Unless the situation or the context could be very clearly open, so that they can see the knowledge that was inputted into the system and not only the summaries it produces, I believe these sorts of systems could cause ethical issues,” he says.
Google DeepMind didn’t explicitly tell participants within the human mediator experiment that an AI system could be generating group opinion statements, even though it indicated on the consent form that algorithms could be involved.
“It’s also vital to acknowledge that the model, in its current form, is proscribed in its capability to handle certain features of real-world deliberation,” Tessler says. “For instance, it doesn’t have the mediation-relevant capacities of fact-checking, staying on topic, or moderating the discourse.”
Determining where and the way this type of technology might be utilized in the longer term would require further research to make sure responsible and secure deployment. The corporate says it has no plans to launch the model publicly.
