AI Use at the DOT to Draft Safety Regulations: Concerns and Implications

The US Department of Transportation's use of AI to draft safety regulations raises concerns about potential errors and safety risks.

AI Use at the DOT to Draft Safety Regulations: Concerns and Implications
United States Department of Transportation

The US Department of Transportation (DOT) has recently adopted artificial intelligence, in particular Google's Gemini model, to draft transportation safety regulations. The decision has raised concerns among staff members that errors in the AI could lead to flawed rules, resulting in injuries or even deaths in the transportation sector.

The Role of AI in Drafting Regulations

The DOT has used Gemini to speed up the regulation drafting process, cutting the time required from weeks to less than 30 days. According to Gregory Zerzan, the DOT's top lawyer, the goal is not to produce a perfect regulation, but one that is "good enough" and can be produced quickly. This approach aims to modernize transportation infrastructure and update outdated rules.

Concerns Within the DOT

Despite the enthusiasm for efficiency, some DOT staff members are skeptical about Gemini's reliability in drafting complex regulations. They stress that drafting regulations requires deep expertise and a detailed understanding of existing laws, a task that AI may not be able to perform with the necessary precision. In addition, AI errors could lead to incomplete or inaccurate regulations, increasing the risk of accidents or harm.

Implications for Transportation Safety

Transportation safety regulations directly affect the safety of airplanes, cars and infrastructure such as pipelines and railroads. Errors in regulations can have serious consequences, such as plane crashes, gas pipeline explosions or derailments of trains carrying hazardous chemicals. Excessive trust in AI without adequate human oversight could compromise public safety.

Conclusion

  • Need for Human Oversight: It is essential that the use of AI in drafting regulations be accompanied by rigorous human oversight to ensure the accuracy and safety of the rules.
  • Risk Assessment: Before deploying AI in critical processes, it is essential to carefully assess the potential risks and implications for public safety.
  • Balancing Efficiency and Safety: While AI can improve efficiency, it must not compromise the quality and safety of regulations, especially in sensitive sectors such as transportation.