You work in AI Safety full-time
But not on strategy. Build a clearer sense of which part of your work is most impactful, so you can do more of it.Know which part of your work matters most.
About this course
Advanced Strategy in AI Safety is for learners who want to think clearly about how advanced AI could reshape society. You will explore how experts forecast AI progress, what could accelerate the development of powerful systems, and why different assumptions lead to radically different futures.
The course examines major challenges such as misalignment, deceptive behavior, misuse, and the concentration of power. You will also consider possible responses, from technical safety and AI control to governance, regulation, and coordination between governments and AI companies.
Rather than asking you to accept a single prediction, the course teaches you to compare competing arguments, question their assumptions, and update your views as new evidence emerges. By the end, you will have a stronger framework for evaluating claims about AI’s future and the decisions that could shape it.
Who is it for:
You want calibration, not vibes
You'd rather hold explicit, testable models of how AI goes than recycle headlines and hot takes.Explicit, testable models, not hot takes.
You've done an intro course
Ours, BlueDot's, or a university group's intro fellowship: take this before going deep on a specific track like governance, alignment, or control.Your step before picking a track.
You're doing a fellowship
Fellowships teach technical skills. We add the complementary strategic reasoning that impresses hiring managers and isn't taught there.Strategic takes fellowships don't teach.
Curriculum
5 UNITS · ~5H EACHNothing to buy. The course builds on Redwood Research’s AI futurism reading list, with added perspectives from other researchers. Every reading is provided online.
Timelines and takeoff
How fast do AI capabilities grow, and what would tell us we're wrong? The models behind short and long timelines, and what takeoff speed changes about everything else.
How experts model the future
The forecasting toolkit behind the field's best predictions: scenario writing, base rates, trend extrapolation, and where each one breaks.
Threat models and key scenarios
The concrete stories of how advanced AI goes wrong, and what each scenario implies about which interventions matter.
The strategic landscape
Labs, governments, and the incentives driving them. What the major actors' plans actually say when you read them closely.
Where the leverage is
Turn your model into a decision: what has to go right after we avoid disaster, which concrete projects are still understaffed, and where your skills and situation can push.
Course Outcomes
An explicit model of the future
Structured, defensible, and yours, not borrowed takes.
Sharper forecasting skills
Forecast timelines and takeoff speeds: when, and how quickly, powerful AIs will arrive.
A read on the strategic landscape
The strategies for preventing AI takeover, and what the labs' and governments' plans actually say.
A peer network in AI safety
People excited about the same questions, and potential future collaborators.
A certificate you can verify
With a public link for your LinkedIn.
A clear next step
A concrete answer to "what do I do on Monday", matched to your skills and situation.
Upcoming cohorts
INTENSIVE
5h/day · 1 week
Dates to be announcedWe’ll publish the next intensive cohort here.
PART-TIME
~8h/week · 3 weeks · twice a week
Dates to be announcedWe’ll publish the next part-time cohort here.
Common questions
Not sure it’s for you? Come to the first session. If it’s not the right fit, we’ll help you find something that is.