Dan Taber, PhD
AI Strategy for High-Stakes Decisions
AI creates a new class of decisions for organizations: when to trust a system, when to slow down, what evidence is sufficient, and how to prepare for failures that haven’t happened yet.
My work sits at the intersection of AI, research, and strategy. I help organizations evaluate emerging AI capabilities, understand their limitations, and build the evidence needed to make better decisions.
Current Focus
My current work explores how organizations evaluate emerging AI systems before established best practices exist. Within Spotify’s Trust & Safety organization, I’m focused on AI evaluation, emerging risks, and the evidence organizations need to make informed decisions about new AI capabilities.
My Approach
I believe the hardest AI challenges are rarely just technical. They emerge where rapidly evolving technology meets imperfect evidence, organizational incentives, and high-consequence decisions. My work focuses on developing practical ways to evaluate AI systems and help organizations make better decisions under uncertainty.
Some of the questions that motivate my work include:
What evidence should leaders require before trusting an AI system?
How do we evaluate capabilities when established benchmarks are no longer enough?
How can organizations anticipate important failures before they reach the real world?