Michael K. Cohen Data-verified

Affiliation confirmed via AI analysis of OpenAlex, ORCID, and web sources.

Researcher

University of Arkansas
Last publication 2026 Last refreshed 2026-05-16

faculty

7 h-index 32 pubs 545 cited

Research Areas

Links

ORCID

OverviewAI-generated summary

Michael K. Cohen's research interests lie in the theoretical aspects of artificial intelligence and advanced artificial agents. His work has explored the risks associated with superintelligent agents and proposed potential pathways for safer AI development, including the role of scientific AI. Cohen has also investigated the intervention of advanced artificial agents in reward systems and the potential for Bayesian oracles to prevent harm. Further research includes the application of algorithmic information theory to understand intelligence and unambitiousness, as well as technical contributions to machine learning, such as log-linear-time Gaussian processes using binary tree kernels and fully general online imitation learning. His publications also touch on detecting cheating in games like chess.

Metrics

  • h-index: 7
  • Publications: 32
  • Citations: 545

Selected Publications

  • Golden Handcuffs make safer AI agents (2026)
    OpenAlex
  • Golden Handcuffs make safer AI agents (2026)
    DOI OpenAlex
  • In Which Areas of Technical AI Safety Could Geopolitical Rivals Cooperate? (2025)
    DOI OpenAlex
  • Imitation learning is probably existentially safe (2025)
    DOI OpenAlex
  • Scientist AI Needs a Government: Structural Governance as the Missing Layer in Non-Agentic AI Safety (2025)
    DOI OpenAlex
  • Superintelligent Agents Pose Catastrophic Risks: Can Scientist AI Offer a Safer Path? (2025)
    10 citations DOI OpenAlex
  • In Which Areas of Technical AI Safety Could Geopolitical Rivals Cooperate? (2025)
    DOI OpenAlex
  • Can a Bayesian Oracle Prevent Harm from an Agent? (2025)
    1 citation DOI OpenAlex
  • RL, but don't do anything I wouldn't do (2024)
    DOI OpenAlex
  • Can a Bayesian Oracle Prevent Harm from an Agent? (2024)
    DOI OpenAlex
  • Regulating advanced artificial agents (2024)
    37 citations DOI OpenAlex
  • Log-Linear-Time Gaussian Processes Using Binary Tree Kernels (2022)
    1 citation DOI OpenAlex
  • Chess: how to spot a potential cheat (2022)
    1 citation DOI OpenAlex
  • Log-Linear-Time Gaussian Processes Using Binary Tree Kernels (2022)
    2 citations DOI OpenAlex
  • Advanced artificial agents intervene in the provision of reward (2022)
    31 citations DOI OpenAlex

View all publications on OpenAlex →