EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

The backfiring effect of weak AI safety regulation

Benjamin Laufer, Jon Kleinberg iD, Hoda Heidari iD

Proceedings of the National Academy of Sciences · 28.07.2026 · 123 (30)

DOI: 10.1073/pnas.2509768123 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the National Academy of Sciences, online erschienen am 20.07.2026. Verfasst von Benjamin Laufer und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Information Science, Cornell Tech, New York, NY 10044. Hauptthemen laut MEDLINE: Artificial Intelligence, Safety. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 50 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Simons Foundation (SF), John D. and Catherine T. MacArthur Foundation (JDCTMF), National Science Foundation (NSF).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Recent policy proposals aim to improve the safety of general-purpose AI, but there is little understanding of the efficacy of different regulatory approaches. We present a strategic model that explores interactions between safety regulation, general-purpose AI technology creators, and domain specialists—those who adapt the technology for specific applications. Our analysis examines how regulatory measures targeting different parts of the AI development chain affect the outcome of this game. Our model assumes AI technology is characterized by two key attributes: safety and performance. The regulator first sets a minimum safety requirement that applies to one or both players. The general-purpose creator then invests in the technology, establishing its initial safety and performance levels. Next, domain specialists refine the AI for their use cases, updating the safety and performance levels and taking the product to market. Resulting revenue is shared between the specialist and generalist. Our analysis reveals two insights: first, weak safety regulation imposed predominantly on domain specialists can backfire. While it might seem logical to regulate AI use cases, our analysis shows that weak regulations targeting domain specialists alone can reduce safety in a large class of parameterizations. Second, in contrast to the previous finding, we observe that stronger, well-placed regulation can mutually benefit all players. When regulators impose appropriate safety standards on both general-purpose AI creators and domain specialists, the regulation can function as a commitment device, leading to safety and performance gains, surpassing what is achievable under no regulation or regulating only one player.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Artificial IntelligenceSafetyHumansModels, Theoretical

Stichwörter der Autor:innen

AI RegulationAI SafetyGame theoryGeneral-purpose AI

Autor:innen und Institutionen

  • Department of Information Science, Cornell Tech, New York, NY 10044.
  • Department of Computer Science and Information Science, Cornell University, Ithaca, NY 14853.
  • Machine Learning Department and the Institute for Software, Systems, and Society, Carnegie Mellon University, Pittsburgh, PA 15213.

ORCID-Profile: Jon Kleinberg, Hoda Heidari

Förderung

  • Simons Foundation (SF)
  • John D. and Catherine T. MacArthur Foundation (JDCTMF)
  • National Science Foundation (NSF)

Zitieren

APA

Laufer, B., Kleinberg, J., & Heidari, H. (2026). The backfiring effect of weak AI safety regulation. Proceedings of the National Academy of Sciences, 123(30). https://doi.org/10.1073/pnas.2509768123

Vancouver

Laufer B, Kleinberg J, Heidari H. The backfiring effect of weak AI safety regulation. Proceedings of the National Academy of Sciences. 2026;123(30). doi:10.1073/pnas.2509768123

BibTeX
@article{Laufer2026,
  author = {Laufer, Benjamin and Kleinberg, Jon and Heidari, Hoda},
  title = {The backfiring effect of weak AI safety regulation},
  journal = {Proceedings of the National Academy of Sciences},
  year = {2026},
  volume = {123},
  number = {30},
  doi = {10.1073/pnas.2509768123}
}
Kostenlos starten →