AI could be used as a tool to strengthen democracy, giving politicians a greater ability to listen to the people who vote for them, according to a leading US expert. 

Speaking at the Ultimo Institute’s first public event, Beth Simone Noveck, an author and world leading expert in democratic innovation and the future of AI, added that technology had evolved to better handle the information available to institutions.

“The real crisis is democratic, not technological,” Noveck said.

“How do we move beyond consultation to give citizens real authority through AI?

“There are very real decisions that we are making today and we need to start talking about what we are using AI for, we have things we need to start doing differently and we need to start demanding this of our institutions.

“We need to start asking institutions to use their buying and regulatory power so that they can act as a countervailing power.”

However, she added government institutions were woefully behind and currently unable to provide this.

Noveck, a professor at Boston’s Northeastern University, who published her 16th book, Reboot: AI and the Race to Save Democracy earlier this year was speaking with Professor Carl Rhodes, director of the newly formed research institute.

Her book is based in part on her experience serving in the White House under Barack Obama at the start of his first term in office in 2009.

In the first 100 days of Obama’s campaign, Noveck worked on a survey asking voters what changes they would like to see, and received almost 200,000 responses. 

“When you give people the ability to talk they have lots to say,” she said. “But that is where that story ends.

“We tried to use technology to allow voters to tell us how to govern but we didn’t have the capability to make sense of all the responses. We gave voters the ability to speak, but our institutions didn’t have the ability to listen. Today we need to get this right.”

We allow too much of the conversation to be taken up by six white guys from Silicon Valley.

Noveck said AI conversations needed to be moved away from tech bros and towards asking whether AI was being used for good or something else. 

“We allow too much of the conversation to be taken up by six white guys from Silicon Valley,” she added. “It is easy to hate Elon Musk, but we need to lobotomise this from the public conversation and consider the stories we are not telling.

“If we don’t have a picture of where we want to head [to], we will go in the wrong direction.”

But she also disparaged anti-tech rhetoric.

Referring to Engaged California, a website run by the State of California, which has developed 19 proposals using public feedback, Noveck suggested AI could be used to help institutions listen better to their constituents. 

“Five of these are now currently in the process of being implemented,” she said. “This is listening, learning and then taking action, consultation processes need to be made more responsive to address a need.

“In regards to the Australian context, all of the ingredients are here, there is a strong democratic culture and lower inequality rates,” she added, citing South Australia’s AI chatbot EdChat, which assists students with school work and teachers with lesson planning, as an example.

Noveck described is as a “tools on models” approach, and that institutions could build models on top of other large language models to make them more transparent.

A 2025 survey by University of Melbourne found 78 per cent of Australians were concerned there would be negative outcomes from AI.

But Noveck said the public needed to put pressure on institutions to use technology to give them greater power and authority over the decisions that will affect them. adding: “The choice is ours, we have to hurry up.” 

Main image of Beth Simone Noveck and Professor Carl Rhodes by Caitlin Young.