by Ryan Cohn
Chief Strategy Officer and Partner, Sachs Media

Two years of survey data show a public that caught up to AI technology faster than it came to trust its government to use it. Here is what the numbers mean for Florida cities, and a practical guide for using AI well.

At the League’s Annual Conference in August, Sachs Media Chief Strategy Officer and Partner Ryan Cohn walked through two years of survey data on how Florida residents view artificial intelligence (AI) use in local government. 68% of Florida voters have now used a generative AI tool themselves, up from 41% just two years earlier. In that same stretch, 68% came to say they do not trust their local government to use AI responsibly. Residents have caught up to the technology faster than they trust local government to use it.

A public that already uses AI

Adoption is close to universal now among Florida residents, and the growth did not come from the youngest residents, who were already on board. The growth came from residents cities often assume they can skip with technology. Among Floridians aged 45 to 64, personal use of generative AI climbed from 41% to 76%, the largest jump of any group. Voters 65 and older more than doubled, from 21% to 50%. The practical lesson is that cities should talk about AI the way you would talk about any tool your constituents already use, because by now most of them do. And cities need to stop writing off audiences by age.

More residents appreciate AI’s uses. 69% now name at least one way AI could benefit local government, up from 61%. Improving efficiency is the benefit people cite most. But the fastest riser is communication: the share who say AI could make a city’s communication with residents clearer grew from 27% to 40%, the biggest move of any benefit in the survey. Efficiency still sells. Clarity is where opinion is shifting, and it is the stronger argument to lead with.

Where the worry has moved

The nature of the public’s concern changed too. Privacy and data security are now the top worries, named by 69% of voters, up from 63%. Bias in AI systems, which led the list in 2024, fell from 66% to 54%. The takeaway is concrete: every public statement about a new AI tool should say plainly how resident data is handled and protected.

A public still making up its mind

For all the movement, most residents do not have a firm view of AI. Asked whether officials should actively use AI, voters gave fewer definite answers on both ends between 2024 and 2026. The firm “yes” slipped from 16% to 12%, the firm “no” dropped from 25% to 15%, and the persuadable middle grew 14 points, to 73%. That is an opportunity for cities to make the case for AI now, before opinions harden, and reach far more people than they may be able to a few years from now.

Underneath the openness is real unease. 63% of Floridians expect humans to lose control of AI at some point, and 57% think it will happen within ten years. That wariness shows up in customer service too: 73% would not be comfortable getting help from an AI chatbot instead of a person when they contact their local government, and only 4% would be very comfortable. Whatever residents do with AI in their own lives, most do not want to be handed to a bot when they need something from their city. Offer AI as an option, and keep a person one step away.

The trust gap, and how to close it

Residents use these tools daily and still doubt their government will use them well. A city closes the trust gap the same way it earns any other kind of trust: by consistently sharing specific messages. Say how your city uses AI. Say just as plainly what AI does not decide. Say how data is protected, and name who is accountable when something goes wrong. The AI Hub’s guidance on communicating about AI builds on exactly these four questions, and cities already publish the tools they use in plain language.

A practical playbook for using AI well

The second half of the conference session focused on what officials actually do with these tools, and where common habits are problematic. The thread running through all of it is simple: use AI for tasks it can handle, and stop asking AI to do the work better done by your team.

A few of the suggested moves work especially well for city tasks. Instead of asking AI to write a statement on a rezoning, where you already know your position, ask it to read the 300 resident emails about the rezoning and summarize residents’ feedback. Instead of a 1,200-word newsletter nobody finishes, ask for a 100-word version a resident who has never been to a meeting could follow. Before a tool answers a policy question from the open internet and gets it wrong, hand it the actual ordinance and tell it to answer only from your city’s specific ordinance. And before you accept a plan AI just praised, ask the technology for the strongest argument against the plan, since AI is built to agree with you.