Politics
Blake Dowling: The AI Paradox needs guardrails, not panic
Artificial intelligence has gone full circle on a flight of panic, heading into yet another batch of turbulence this month.
A couple of years ago, we moved from serious concerns to complete panic, then to rapid user adoption and massive productivity spikes, then back to serious concerns and complete panic.
I called this the “AI Paradox” in a WCTV report.
Without further delay, welcome to the AI Paradox: A place where you can hold several thoughts and opinions, all of which make sense.
For example: Slow down.
Make sure security and guardrails are in place. Avoid another AI swarm, like the situation involving Hugging Face, where AI bots reportedly went beyond their assigned tasks, accessed the internet, communicated without our knowledge, and ventured well outside the lines.
That is smart. So is outpacing China and other global powers so America leads the race, much like the race to the moon or the arms race.
Both instincts are smart.
Last week, the full-on panic announcement from AI corporate leaders caught the world off guard. Their warning that this is all very bad left many of us scratching our heads: You got us here, right? In the early days, AI skeptics and critics said many of the same things these leaders now shared. What are they really up to? Are they trying to keep smaller AI companies out?
Was this month’s messaging strategic panic aimed at creating barriers?
In Washington, the AI Paradox is on full display, as expected. Ideas range from banning AI to accelerating it. No surprises there. Hear both sides at the pro-human conference.
From the Microsoft camp came one cool move. While others drew lines in the sand, ranted, raved and panicked, Microsoft put out a calm announcement about an AI code of conduct.
What are the real fears? The AI headline about biological weapons looks like media smoke and mirrors: Someone asked AI how to make a bioweapon, and the AI reported it. The real fears are AI swarms, AI in the hands of hackers and rogue nations, and the possibility of AI building the next generation of AI.
The real threats in the Anthropic threat intelligence report are right out of Hollywood: AI taking over weapons systems, surveillance, cybercrime and hacking and phishing on steroids.
What does hacking on steroids look like? Imagine a hacker registers a fake domain for a dentist’s office in Missouri and sends emails asking for money or passwords, cleverly disguised as messages from someone the recipient knows, such as a software vendor, bank, or government agency. That attack takes time. With an AI tool that does not recognize the behavior as bad, the hacker could target an entire state, every dentist and thousands of inboxes at once. That is AI hacking on steroids.
The Anthropic report also raises one other item needing attention: recursive self-improvement. That means AI learning on its own and creating the next generation of thinking machines. Some experts say this is the real item to fear; others say it is the goal.
Here in Tallahassee, we decided to bring the brightest minds from Florida State University together with private-sector leaders — THF, Sachs, Phusion AI, Ruvos and more — plus national partners Kaseya and Sherweb for a real AI discussion.
On Oct. 9, we will separate panic from progress and talk about using AI safely in Florida at the first annual Aegis AI Summit and Workshop.
We cannot set national AI policy, but we can help local companies protect their data with guardrails and showcase Florida innovation in AI utilization.
Bottom line, folks: There will always be evil in the world.
Those seeking to harm others will use every available tool. Criminals, hackers, terrorists — you name it. AI is another tool in the arsenal, and we have to fight back by examining sketchy emails, texts and phone calls before giving away credentials.
Businesses should deploy every available cyber tool, especially those that prevent executable malware from launching when someone clicks the wrong thing or when a hacker gets into the network another way.
Flight 2026 has had plenty of turbulence, especially in September. Isn’t there a song about that? Wake Me Up.
Maybe we will hit cruising altitude in the Fall. For now, keep your seat belts fully fastened.
The post Blake Dowling: The AI Paradox needs guardrails, not panic appeared first on Florida Politics – Campaigns & Elections. Lobbying & Government..