Florida lawmakers face a two-sided challenge on artificial intelligence: protecting residents from its clearest dangers without imposing broad or unstable rules that stifle innovation.
That’s according to panelists at the Build American AI panel discussion in Tampa, moderated by Florida Politics publisher Peter Schorsch.
The panelists agreed that guardrails are necessary, particularly around deepfakes, child sexual abuse material, deceptive political content and companion chatbots. But they also urged lawmakers to keep the rules predictable and focused on demonstrated harms while ensuring Florida has the infrastructure needed to compete in an AI-driven economy.
Here are six of their main recommendations:
Keep the rules simple and stable
Adam Basford, Vice President of Government Affairs at Associated Industries of Florida, said businesses are accustomed to complying with regulations.
But he said a greater concern is uncertainty over what the rules will be from one year to the next, and advised regulators to avoid moving goalposts for industries looking to innovate while complying with any potential state or federal laws.
“Companies aren’t scared of regulation,” Basford said. “They understand how to deal with regulations. They understand how to comply with what’s out there. What they’re scared of is inconsistency. What they’re scared of is a consistently changing opportunity there.”
Basford also said government should continue to regulate the “low-hanging fruit.”
“We dealt with CSAM (Child Sexual Abuse Material). We’ve dealt with deepfakes and political advertisements. We’re dealing with those kinds of things,” he said. “There are other things out there dealing with child protection with companion chat bots. Those are obvious things that there needs to be some regulatory process around.”
Distinguish between different AI uses
Basford said lawmakers should also avoid placing fundamentally different applications of AI under the same regulatory framework simply because they use similar technology.
He pointed to companion chatbots used by children as an area that raises legitimate concerns about private conversations and potential harm. But he argued those risks are different from those presented by a business.
“We saw simple things in the discussion about the bill of rights dealing with companion chatbots, and that’s obviously something dealing with our children, behind closed doors and in personal situations,” Basford said.
“There should absolutely be some thought about how we handle and protect that. But that was kind of lumped into the same discussion as how our theme parks utilize AI in forward-facing character interactions. Those are totally different things and need to be handled differently.”
Avoid conflicting state standards
Build American AI Director Nathan Leamer supports a national baseline for issues that routinely cross state borders, including cybersecurity, transparency and harmful deepfakes.
He argued that online products, data and AI-generated content do not remain within one jurisdiction. A system in which states establish substantially different requirements could leave businesses attempting to determine which standard applies whenever a product reaches users across state lines.
“I think Ron DeSantis is correct that there is a role for states to play. But it’s not to be creating a patchwork of different rules that companies have to jump through the hoops to get there. It’s more about creating certainty and enabling. For example, schools. How do schools handle AI? It’s a question for the states, in many ways, that’s not a federal thing,” Leamer said. “That’s the kind of balance that we’re looking for.”
He said states also need to make decisions about how they choose to handle data centers and energy production, criticizing New York for taking a stance on data centers only after also shutting down a nuclear facility.
“New York is making a decision to say we’re a no to data centers. They also made a decision to shut down a nuclear facility that actually was generating a lot of power a couple of years ago. That has hindered their ability to compete. Other states, they have the power and resources to compete,” Leamer said.
Give consumers appropriate transparency
Audra Nasser, Founder of Dinner Done, said transparency will be critical to preserving public trust as businesses adopt AI, particularly for small companies whose reputations depend heavily on relationships with their communities.
Nasser said she became enthusiastic about the potential boost that artificial intelligence gives business owners after first interacting with ChatGPT to assist with social media copy. From there, her interest grew, however she also said that she experienced resistance from more apprehensive employees who felt threatened by embrace of the technology.
Nasser explained that an intern expressed concern not only for the impact on workers, but also for perceptions among potential customers or members of the public who outright reject use of AI.
Based on the feedback, Nasser said she has embraced use of AI but is mindful of customer trust. She draws the line at generative artwork and video, but does use it through other means for the business. She explained she may be willing to experiment further in the future, once general perceptions about AI use become more accustomed to the emergent technology.
Nasser said establishing clear expectations and disclosure could give responsible businesses a way to use the technology without making customers suspicious of everything they produce.
“We’re a small business. We live and die by our community and we’ve spent 20 years building trust. So I have some red lines.”
Preserve practical uses
Prasad Chittaluru, CEO of EPIC Engineering & Consulting Group, said lawmakers and government agencies should avoid policies that prevent public employees from using AI for routine administrative and infrastructure-management tasks, suggesting that it should also help streamline some of government’s most dubious decisions.
He said government agencies often manage roads, water systems, sewer systems, airports and other public assets through separate departments that do not always share information effectively. That lack of coordination can produce familiar inefficiencies, such as one agency repaving a road only for another to tear it up months later to complete separate work.
Better-managed data could allow agencies to identify overlapping projects, coordinate schedules and explain their spending decisions more clearly to elected officials and residents.
“Transparency is the key,” Chittaluru said. “When people can see what is going on, they can better react.”
He argued AI should be used to reduce repetitive work such as tasks that involve reviewing data and producing routine reports, freeing public employees to spend more time making decisions and communicating with residents.
The goal, he said, is not to eliminate public-sector jobs or give agencies a tool to monitor workers. It is to provide employees with better information and allow them to focus on work requiring human judgment rather than repetitive data processing.
“Whether you like it or not, AI is coming,” Chittaluru said. “Let AI help get rid of the mundane stuff. Get it off the daily crunch of the report. Let that automate. Use your time to make your decisions faster.”
Plan for broadband, power and data centers
Panelists argued that policy debates should also include the physical infrastructure required to support AI, such as basic internet access for underserved communities.
Leamer said broadband access could create winners and losers based on which Floridians can use emerging tools and benefit from the economic opportunities they create. He argued residents should not be excluded simply because internet service is less reliable in their communities.
“If we want every American, every Floridian and everyone in this country to be able to have access to AI, we need to make sure they have good internet access and the ability to have these tools at their disposal. You should not be redlined based on where you are,” Leamer said.
Julio Fuentes, President and CEO of the Florida Hispanic Chamber, said data center advocates also need to do more to communicate with the communities surrounding proposed facilities.
Fuentes has publicly supported data centers as important economic infrastructure, but he acknowledged experiencing his own moment of doubt when Project Tango was proposed near his West Palm Beach-area community.
“Literally it hit me like, ‘Whoa, wait a second. I like these things, but I don’t know about my own backyard.’ I had that kind of kick back for one second, because I know the details behind these data centers,” Fuentes said.
“But I think that the job has not been done in talking to local communities, parents, schools, business leaders about how these data centers actually work, and what it’s actually going to do to actually help your back yard, not damage it.”
Opposition to Project Tango became too intense to overcome, and the proposal was rejected this week. Fuentes said the experience showed that developers cannot wait until public resistance has hardened before explaining a project.
“Yesterday it actually got voted down, and simply because the outrage from the community was overwhelming. There’s no coming back from that backlash,” Fuentes said.
“I think moving forward, we need to do a lot more education on what these data centers are, and what they can do for our own communities. We’re talking about new firehouses, new schools being built, teachers could get paid what they deserve to get paid. We see this happening in other parts of the country. We just need to do a better job here.”