Collaborative Research: EAGER: FDASS: Towards An American Alternative to the European AI Regulation

Today, Europe is leading the way in developing laws to manage the increasing range of harm that artificial intelligence (AI) can and does create. Europe's AI Act focuses on regulating how its businesses develop and use AI. Many experts worry that this heavy-handed intervention into private enterprise will stifle AI innovation, including, ironically, innovations that could ultimately make European citizens safer. The United States urgently needs its own legal approach for managing AI harms. This project's novelty lies in developing an objective standard for deciding who is at fault after an AI harm occurs. This approach avoids Europe's tussle between law and computing because an after-the-harm standard does not directly interfere with business operations. Rather, it provides businesses the space to innovate and the incentive to figure out how best to design accountable software systems that minimize avoidable AI harms.

The objective of this project is to bring accountability for AI without impeding the businesses' and researchers' ability to continually innovate and lead in AI. To do so, this project aims to design an objective-fault standard for AI that does not prohibit or censure any AI behavior outright but instead compares AI’s behavior with an external negligence benchmark. Then, by calibrating the benchmark’s standard to the social value and the current safety profile of the AI conduct at issue, the AI law could be applied flexibly, progressively, and across broad domains. The researchers plan to attain this objective by (i) laying the legal foundation for a negligence standard for AI, (ii) developing AI negligence benchmarks for three representative applications, and (iii) evaluating this new standard against the European AI Act. Upon completion, this research would support establishing a distinctly American alternative to the European style AI regulation.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

Award Number: 2532589
Principal Investigator: Supreeth Shastri
Funds Obligated: $148,341
State: TX