[ Question ]

How can I help research Friendly AI?

by avichapman 2mo9th Jul 20193 comments

23


I've just started my journey toward a PhD in machine learning. My first steps are in computer vision. But my long term goal is to be able to help in the effort to solve the alignment problem and produce Friendly AI before any non-friendly alternatives emerge.

What sort of trajectory should I aim for with my research? What sort of post-PhD jobs should I be aiming for? Who should I be making contacts with as I go?

New Answer
Ask Related Question
New Comment
Write here. Select text for formatting options.
We support LaTeX: Cmd-4 for inline, Cmd-M for block-level (Ctrl on Windows).
You can switch between rich text and markdown in your user settings.

2 Answers

Hi. I'm also a PhD student (upcoming fourth year). Last year, I wrote about my process of auto didacting and switching research areas to directly work on FAI. Beyond that advice, I have more which is yet unpublished: select your committee wisely. I purposefully chose those professors who were easy to get along with and who seemed likely to be receptive to longer-term concerns.

I generally recommend working on an alignment research area in your free time, levelling up as necessary. One path I took: make some proposals and demonstrate you can think novel thoughts and do research. Then, get funding to make this your full-time research.

Alternatively, you can just level up while in your program. Check out Critch's Deliberate Grad School and Leveraging Academia.

If you're interested in a Skype sometime, feel free to message me. There's a MIRIx Discord server I can invite you to as well.

Welcome to the journey!

If you haven't already, check out the 80,000 Hours website (their goal is to provide useful advice on how people can use their career to do the most good).

Here are some links that seem relevant specifically for you (some might be out of date):
https://80000hours.org/key-ideas/ (see "AI safety technical researcher" box)
https://80000hours.org/articles/high-impact-careers/#ai-safety-technical-researcher
https://80000hours.org/problem-profiles/positively-shaping-artificial-intelligence/
https://80000hours.org/career-reviews/artificial-intelligence-risk-research/
https://80000hours.org/career-reviews/machine-learning-phd/

You can also apply for their coaching service.