These articles are our retrospective on our [AI alignment course](https://aisafetyfundamentals.com/alignment/) iteration, beginning in March 2024.[1] We identified different perspectives we'd like to analyse the course from, and broke up our analysis into three parts:

- [Systems and processes](https://bluedot.org/blog/ai-alignment-march-2024-retro-systems-and-processes/)
- [Curriculum development](https://bluedot.org/blog/ai-alignment-march-2024-retro-curriculum-development/)
- [Getting good people on the course](https://bluedot.org/blog/ai-alignment-march-2024-retro-good-people/)

<Embed url="https://storage.k8s.bluedot.org/website-assets/migrated/e3ca7376-8c56-472f-abd0-83a33238d8cf.svg" />

## Who is this for?

It's primarily written for ourselves, and published in line with our philosophy of [working in public](https://bluedot.org/blog/working-in-public/). However, we expect it could be useful to:

- People considering [applying to our courses](https://apply.aisafetyfundamentals.com/alignment), to understand the kind of work we put in to make them an excellent experience (and what the experience is like for participants).
- People considering [joining our team](https://bluedot.org/careers/), to better understand the kind of work we do.
- People [running local versions of our courses](https://bluedot.org/running-versions-of-our-courses/), or otherwise getting into course design or running, to copy what we did well and avoid the mistakes we made.
- People [considering giving us funding](https://bluedot.org/blog/case-for-support-2024-2026/), to help explain how we might use this funding to deliver on supporting people developing knowledge, skills and connections to pursue high-impact careers.

## Footnotes

[1]: This was also previously referred to as the Early 2024 AI alignment course, or April 2024 AI alignment course. We have since standardised our course iteration naming.


What we learnt from running our AI alignment course in March 2024

Adam Jones • July 2, 2024