Transportation Department Might Use AI To Write Rules In Minutes, And Some Say That’s Terrifying
The Department of Transportation plans to utilize artificial intelligence to draft regulations, aiming to streamline a process that traditionally takes months or years.
Reports indicate that Google Gemini could potentially create these regulations in mere minutes, raising significant concerns regarding the quality and safety of such hastily produced rules.
The Speed vs. Quality Dilemma
The push for AI-generated rules has triggered alarm bells, especially given a government lawyer’s assertion that “we don’t need perfect rules, just ones that are good enough.” This sentiment suggests a troubling prioritization of quantity over quality, particularly within an agency responsible for public safety.
According to reports from ProPublica, the DOT’s general counsel, Gregory Zerzan, emphasized a need for speed, stating, “We want good enough. We’re flooding the zone.” Such statements imply a willingness to produce a large volume of regulations, potentially compromising safety standards.
AI’s Role in Regulatory Drafting
The motivation behind employing AI like Gemini is clear: the drafting and revising of federal regulations can be time-consuming. In stark contrast, this AI could reduce that time to as little as 20 minutes for a draft, facilitating a rapid transition from proposal to a complete draft ready for review in just 30 days.
Notably, a presentation suggested that AI could handle approximately 80% to 90% of the regulatory writing tasks, leaving humans to oversee the remaining portion. This raises critical questions about the adequacy of human oversight in identifying potential hallucinations and errors in AI-generated content.
Concerns from Industry Experts
Despite the potential efficiency gains, many employees and former officials express concerns about this shift. Mike Horton, a key figure in the discussion, likened the use of Gemini for rulemaking to “having a high school intern.” Such comparisons underline the skepticism regarding the capability of AI to produce regulations that meet rigorous standards.
The Road Ahead
As this initiative unfolds, the effectiveness of using AI in regulatory processes remains to be seen. It will depend heavily on human involvement in the oversight of AI-generated content to ensure that rules are not only efficient but also safe and compliant with established practices.
Ultimately, the aim for “good enough” may not provide the reassurance that both officials and the public seek, leaving a cloud of uncertainty over the future of regulatory practices in the age of AI.
