Can we balance efficiency with human experience in civic engagement with AI on the horizon

Though AI, including large language models (LLMs) like ChatGPT, may still seem fairly new to us, we're already seeing the ways it'll have huge impacts on our lives. Urban Planning Professor Pamela Robinson is interested in what happens when AI is used in her field.

"I was always interested in weird things from the future and what they'd mean for Canadian cities," she says. Her latest report, Artificial Intelligence: Its potential and ethics in the practice of public participation, with Waterloo PhD candidate Morgan Boyco, dives into ways AI can be used in urban planning/community engagement processes, and the benefits and concerns.

The paper, commissioned by the International Association for Public Participation Canada, explores ways in which public participation practitioners (including city planners) might encounter AI, and how they might leverage the technology to enhance democratic processes.

If technology is used in civic conversations, it should support people, not replace them," says Robinson. "That's really important, because city building depends on learning from people's lived experiences, we need to listen carefully for things that are unique or new." Using an LLM to filter through responses and determine what the average answer is in these participatory projects will flatten that human experience. "The full diversity of people's opinions and ideas are critical in city building processes," she says.

"The report has all these potential AI encounters at different stages of a participatory process, but we also wanted to equip people to think about the ethical implications, which aren't always obvious," says Boyco. "We wanted those working in public participation to consider how decisions are made collectively and how AI might impact that."

Thinking about how AI will be used in any field is top of mind for many professionals right now, and public participation professionals need the same primer. The report outlines several ways AI can be used in the field, including for interaction and dialogue, for planning and brainstorming sessions, and for results analysis support - but each use case is not without its concerns.

"I think that there are significant questions about what happens when we move from using people to summarize comments and input into running those through an LLM," says Robinson. "We as planners learn by reading people's contributions firsthand. If we turn over the data to technology and don't engage with it ourselves, we won't know what we're missing."

Robinson says transparency and accountability will be important as AI becomes more deeply embedded in our civic processes. "We want people to bring their human selves to the process, not their large language model interpretation of what they think matters," she says. "If we start taking humans out of the process, it's not going to help with trust between residents and government decision-makers. There's the potential to develop a vicious cycle that we talk about in the report, as 'synthetic democracy'."

Examining how humans could be left out of the participatory process is one major challenge AI presents to the field. But which humans are left out is another. The data used to train these machines remains a blackbox. "Whose data trains the machine?" is an important question, says Robinson. "How representative are the training data of the communities we work with? City building is place-based work so bringing technology tools that might flatten the uniqueness of our communities is a problem." Is AI defaulting to generic data sets it's been trained on, or will it honour who our community members are?

The report includes potential benefits to the technology as well, and Robinson says academic institutions like TMU have a real opportunity to explore AI with students. "I wonder: will it free up time for us to have richer conversations about the things that really matter? These tools promise efficiency but it isn't the only goal in city building - we want a public transit system that works efficiently. But if you think about our most treasured places in the city - Kensington Market,Scarborough - these places are vibrant and curious. We need to be careful about just getting stuck in the optimization focus."

Read Robinson and Boyco's report, which they'll present at the International Association of Public Participation (IAP2) in October.

Public link

Please use the above public link if you want to share this noodl on another website.