The Department of Defense might be the first government agency to roll out a department-wide AI chatbot, but the Department of Transportation is about to be the first to draft actual binding regulations with the tech.
According to a new investigation by ProPublica, the top transportation agency has tapped Google Gemini to help write new regulations affecting aviation, automotive, railroad, and maritime safety. In internal communications from DoT attorney Daniel Cohen, agency staffers were presented with the plan along with a demonstration of AI’s “potential to revolutionize the way we draft rulemakings.”
The AI demonstration, Cohen enthused, would show off “exciting new AI tools available to DOT rule writers to help us do our job better and faster.”
His focus, disturbingly, was explicitly that AI is fast, even if it isn’t particularly accurate.
“We don’t need the perfect rule on XYZ. We don’t even need a very good rule on XYZ,” DoT general counsel Gregory Zerzan said, according to the recent meeting notes obtained by ProPublica. “We want good enough,” he said. “We’re flooding the zone.”
In addition, Zerzan noted that enthusiasm for the DoT AI tool runs all the way to the top, noting that Donald Trump is “very excited about this initiative.”
Six DoT workers who spoke to ProPublica anonymously said that typical regulation-writing can take months, and sometimes years, due to the complications involved. But at the demonstration in December, a presenter told them Google’s Gemini could cut that down to “minutes or even seconds.”
Zerzan, for his part, told DoT staffers that the goal is to be able to pump out a new regulation in as little as 30 days. “It shouldn’t take you more than 20 minutes to get a draft rule out of Gemini,” he told regulators.
Asked for his opinion by ProPublica, the DoT’s former chief AI officer Mike Horton compared the plan to “having a high school intern that’s doing your rulemaking.”
For anyone worried about keeping trains on the tracks and planes in the sky, it’s an incredibly troubling development. Large language models (LLMs) like Gemini are prone to errors known as hallucinations. Gemini has been linked to a number of embarrassing episodes, like hallucinating marriages that don’t exist or making up dangerous medical misinformation.
And that’s without getting into how, in April of last year, Gemini users were perturbed as the chatbot abruptly started spitting out screeds that read like disturbing psychological breaks — the last thing you want when dealing with federal regulations on activities as crucial as air traffic control.
More on Trump: Trump’s HHS Trashes Top African Health Organization as “Fake” and “Powerless”
The post Trump Department Responsible for Airline Safety Using AI to Write New Regulations, So They Can Be Churned Out as Fast as Possible appeared first on Futurism.
