AI alignment
The enormous problem facing humanity.
May 14, 2023
My views on the importance and difficulty of the AI alignment problem.
Tim Bakker
Senior ML Researcher at Qualcomm AI
My current research interests focus on AI safety, in particular CoT and latent space monitoring. I completed my PhD in machine learning at the University of Amsterdam, supervised by Max Welling and Herke van Hoof.
I am concerned about the potentially transformative risks associated with advanced AI and have spoken at events including the Dutch National AI Debate, Pakhuis de Zwijger, the Dutch Ministry of Defence, and University of Amsterdam. I have also written for the NRC newspaper on the approaching AI revolution.
Featured Publications
View all →Monitorability Is Trainable: Post-Training for Chain-of-Thought Monitorability
Usman Anwar*, Tim Bakker*, Christos Louizos, David Krueger
* Equal contribution
arXiv preprint, 2026. Under submission at ICLR, 2027
Relative Representations Enable Portable White-Box Monitoring Across Language Models
Tim Bakker, Christos Louizos
arXiv preprint, 2026. Under submission at ICLR, 2027
Analyzing and Improving Chain-of-Thought Monitorability Through Information Theory
Usman Anwar*, Tim Bakker*, Dana Kianfar, Cristina Pinneri, Christos Louizos
* Equal contribution
arXiv preprint, 2026
Experimental design for MRI by greedy policy search
Tim Bakker, Herke van Hoof, Max Welling
NeurIPS, 2020. Spotlight
Featured Posts
View all → 
From Bayesian inference to modern machine learning
How Bayesian inference is hiding in your non-Bayesian models
April 6, 2021
A tour from prescriptive Bayesianism to modern statistical machine learning.
Featured Public Engagement
View all →
Nederland is niet voorbereid op de risico's van de rap naderende AI-revolutie
February 7, 2025
Article in the national Dutch newspaper NRC, on current AI developments and the importance of global action for mitigating risks to the economy, risks of arms races, and risks of loss-of-control, by Otto Barten and Tim Bakker.
Panel on the Existential Risks of AI with prof. Stuart Russell
Panel at Pakhuis de Zwijger · Pakhuis de Zwijger, Amsterdam
April 12, 2023
Panel on the existential risks of artificial intelligence, with Stuart Russell, Maxim Februari, Queeny-Aimee Rajkowski, Tim Bakker, and others.