Dr. Kevin Baum

I’m a philosopher and computer scientist working on AI alignment and human-centered control. I ask

what norms AI agents should answer to, who gets to decide (i.e., questions of normative authority), and how the relevant normative expectations can be harvested in practice and represented technically,
how to build agents that behave aligned to these norms — or at least follow those norms — whether or not they expect to be watched, and how this can be realized in a contestable and iteratively improvable way,
and how humans keep effective oversight and control in a way that is safe and meaningfully preserves moral responsibility and human autonomy.

Where to start: Neither Dogma nor Taste (normative authority) · GRACE (agents that follow norms) · On the Quest for Effectiveness in Human Oversight (oversight and control)

→ More about my research · Essays

Current positions:

Independent Research Group Leader, Responsible AI and Machine Ethics (RAIME) at the German Research Center for Artificial Intelligence (DFKI)
Saarland University Associate Fellow, Saarland Informatics Campus (with PhD supervision rights in Computer Science)
Research Associate, Oxford Internet Institute, University of Oxford
Senior Researcher, Institute for Ethics in Technology, Hamburg University of Technology

I also serve on the executive board of the Center for European Research in Trusted AI (CERTAIN) and am a member of the International Association for Safe and Ethical AI (IASEAI), founded by Stuart Russell.

I co-founded Algoright e.V. and advise on AI policy → Engagement

selected publications

IASEAI

Breaking Up with Normatively Monolithic Agency with GRACE: A Reason-Based Neuro-Symbolic Architecture for Safe and Ethical AI Alignment

Felix Jahn, Yannic Muskalla, Lisa Dargasz, and 2 more authors

In Proceedings of the 2nd Conference of the International Association for Safe and Ethical Artificial Intelligence (IASEAI 2026), 2026

Bib

@inproceedings{jahn2025breaking,
  author = {Jahn, Felix and Muskalla, Yannic and Dargasz, Lisa and Schramowski, Patrick and Baum, Kevin},
  title = {{Breaking Up with Normatively Monolithic Agency with GRACE: A Reason-Based Neuro-Symbolic Architecture for Safe and Ethical AI Alignment}},
  year = {2026},
  booktitle = {Proceedings of the 2nd Conference of the International Association for Safe and Ethical Artificial Intelligence (IASEAI 2026)},
  volume = {2},
  pages = {254--267},
  doi = {10.1609/iaseai.v2i1.43029},
  url = {https://ojs.aaai.org/index.php/IASEAI/article/view/43029},
}

NeurIPS-W

Norms at a Price: Why RL-Based Alignment Can Promise Conditional Compliance at Best

Kevin Baum, Rūta Binkytė, and Felix Jahn

2026

Foundations of Agentic Systems Theory (FAST) Workshop at NeurIPS 2026 (accepted full paper); preprint arXiv:2609.07627

Bib

@misc{BaumBinkyteJahn2026,
  author = {Baum, Kevin and Binkytė, Rūta and Jahn, Felix},
  title = {{Norms at a Price: Why RL-Based Alignment Can Promise Conditional Compliance at Best}},
  year = {2026},
  note = {Foundations of Agentic Systems Theory (FAST) Workshop at NeurIPS 2026 (accepted full paper); preprint arXiv:2609.07627},
  doi = {10.48550/arXiv.2609.07627},
}

LNCS

Justifications for Democratizing AI Alignment and Their Prospects

Andre Steingrüber, and Kevin Baum

In Third International Conference on Bridging the Gap between AI and Reality: AISoLA 2025, Rhodes, Greece, November 1–5, 2025, Selected Papers, 2026

On-site proceedings of the 3rd AISoLA Symposium, Rhodes, Greece, Nov 1–5, 2025

Bib

@inproceedings{SteingrueberBaum2026,
  author = {Steingr{\"u}ber, Andre and Baum, Kevin},
  editor = {Steffen, Bernhard},
  title = {{Justifications for Democratizing {AI} Alignment and Their Prospects}},
  booktitle = {Third International Conference on Bridging the Gap between {AI} and Reality: {AISoLA} 2025, Rhodes, Greece, November 1--5, 2025, Selected Papers},
  series = {Lecture Notes in Computer Science},
  volume = {16220},
  pages = {146--159},
  publisher = {Springer Nature Switzerland},
  address = {Cham},
  year = {2026},
  doi = {10.1007/978-3-032-07132-3_10},
  note = {On-site proceedings of the 3rd {AISoLA} Symposium, Rhodes, Greece, Nov 1--5, 2025},
}

AIES

Aggregation Problems in Machine Ethics and AI Alignment

Kevin Baum, and Marija Slavkovik

In Proceedings of the 8th AAAI/ACM Conference on AI, Ethics, and Society (AIES 2025), Madrid, Spain, Oct 20–22, 2025, 2025

Bib

@inproceedings{BaumSlavkovik2025,
  author = {Baum, Kevin and Slavkovik, Marija},
  title = {{Aggregation Problems in Machine Ethics and AI Alignment}},
  booktitle = {Proceedings of the 8th AAAI/ACM Conference on AI, Ethics, and Society (AIES 2025), Madrid, Spain, Oct 20--22, 2025},
  year = {2025},
  address = {Madrid, Spain},
  publisher = {ACM},
  doi = {10.1609/aies.v8i1.36554},
}

FAccT

On the Quest for Effectiveness in Human Oversight: Interdisciplinary Perspectives

Sarah Sterz, Kevin Baum, Sebastian Biewer, and 4 more authors

In , Jun 2024

Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency (FAccT 2024)

Bib

@inproceedings{sterz2024,
  author = {Sterz, Sarah and Baum, Kevin and Biewer, Sebastian and Hermanns, Holger and Lauber-Rönsberg, Anne and Meinel, Philip and Langer, Markus},
  title = {{On the Quest for Effectiveness in Human Oversight: Interdisciplinary Perspectives}},
  month = jun,
  year = {2024},
  note = {Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency (FAccT 2024)},
  doi = {10.1145/3630106.3659051},
}

M&M

Effective Human Oversight of AI-Based Systems: A Signal Detection Perspective on the Detection of Inaccurate and Unfair Outputs

Markus Langer, Kevin Baum, and Nadine Schlicker

Minds and Machines, Jun 2024

Bib

@article{Langer2024-LANEHO-2,
  author = {Langer, Markus and Baum, Kevin and Schlicker, Nadine},
  doi = {10.1007/s11023-024-09701-0},
  journal = {Minds and Machines},
  number = {1},
  pages = {1--30},
  title = {{Effective Human Oversight of AI-Based Systems: A Signal Detection Perspective on the Detection of Inaccurate and Unfair Outputs}},
  volume = {35},
  year = {2024},
}