Roman V. Yampolskiy

About

Roman V. Yampolskiy is a pioneering researcher at the intersection of artificial intelligence safety, machine ethics, and biometric identification of artificial entities. Best known for his foundational work in AI safety engineering, Yampolskiy has challenged mainstream assumptions in the field — most notably arguing in his highly cited 2012 paper (122 citations) that machine ethics represents a fundamentally misguided approach to ensuring AI safety, instead advocating for rigorous safety engineering frameworks tailored to artificial general intelligence. This position, elaborated across multiple works totaling hundreds of citations, has made him a distinctive and provocative voice in the AI safety debate. Yampolskiy has also made significant contributions to understanding how dangerous AI systems emerge, developing taxonomies of pathways to dangerous AI that move beyond science-fiction tropes toward technically grounded risk analysis. His collaborative work on moral autonomous systems bridges the often-separate worlds of philosophy and engineering. Perhaps uniquely, he pioneered the concept of "artimetrics" — applying biometric principles to authenticate robots, avatars, and intelligent software agents — a forward-thinking research thread that anticipated today's concerns about verifying artificial identities. Across more than a decade of prolific scholarship, Yampolskiy has established himself as one of the field's most original thinkers on existential risks posed by advanced AI.

Research Focus

Key Achievements

Top Papers

1. 1
   Artificial Intelligence Safety Engineering: Why Machine Ethics Is a Wrong Approach

   122 citations · 2012
2. 2
   Safety Engineering for Artificial General Intelligence

   76 citations · 2012
3. 3
   Towards Moral Autonomous Systems

   47 citations · 2017
4. 4
   Taxonomy of Pathways to Dangerous Artificial Intelligence

   41 citations · 2016
5. 5
   Artimetrics: Biometrics for Artificial Entities

   40 citations · 2012
6. 6
   Applying Biometric Principles to Avatar Recognition

   39 citations · 2010
7. 7
   Applying Biometric Principles to Avatar Recognition

   39 citations · 2011
8. 8
   What to Do with the Singularity Paradox?

   29 citations · 2012
9. 9
   Taxonomy of Pathways to Dangerous AI

   11 citations · 2015
10. 10
    State-of-the-Art in Robot Authentication [From the Guest Editors

    10 citations · 2010

Key Collaborators

Contact & Links