AI safety fieldbuilding in Warsaw, Poland (funding for 1 semester)

Project summary

Some of the most talented AI people in the world are Poles. 10 out of 50 early OpenAI employees were Polish (source: https://therecursive.com/openai-in-poland-fireside-chat-with-wojciech-zaremba-sam-altman-and-szymon-sidor/ ). Most of them come from just one elite university: University of Warsaw. Remarkably, not only isn't there any AI safety org trying to reach its students. There isn't a single AI safety org in the whole Poland.

What are this project's goals? How will you achieve them?

Reaching out to Poland's best technical students and convincing them that AI safety is a thing seems like a neglected, impactful and tractable cause area. Furthermore, it seems that progress could be made quite inexpensively.

Concretely: the project would provide the first exposure to AI safety for skilled technical students. Ideally, we would help the students gain enough knowledge and expertise to be a good fit for programs ranging in difficulty from ML4Good to MATS.

To be clear, by AI safety I mean decreasing x-risks and s-risks posed by a future unaligned artificial superintelligence. There are many paths toward that noble goal. I plan to focus my fieldbuilding efforts mainly on the technical route, and to a lesser extent on outreach and policy. I currently regard a complete moratorium / ban on frontier AI model development as completely unrealistic, so I will not advocate for it.

The funding will be for 1 semester of fieldbuilding.

How will this funding be used?

Organizing events: presentations, hackathons, networking, 1 on 1 mentoring, group projects
Creating a university group with regular meetings
Advertising costs, both offline and online
Travel & accommodation costs for presenters from outside Warsaw to come and present on campus
Optionally: some swag (stickers etc.), pizza & soft drinks for events to help win goodwill of the students
Significant room for scaling up the project with more time and funding: reaching students from relevant but less strictly technical programs (e.g. cognitive neuroscience), reaching other good universities in the area, "preaching" AI safety at tech meetups, running the project for longer.

Who is on your team? What's your track record on similar projects?

Most probable collaborators:

Poles I met at ML4Good: Mikołaj Kniejski, Jakub Nowak, Michał Skowronek
Active members of EA Poland who are based in Warsaw (~15 people)
Active members of Polish online AI Safety reading group

Most probable consultants for the project:

Charbel-Raphael Segerie (CeSIA), Nia Gardner (ML4Good organiser), Carolina Oliveira (operations, Condor Camp)
Experienced EA university group leaders and fieldbuilders such as Chris Szulc (leader of EA Poland)
Polish AI safety "rockstars" involved with EA, such as Tomek Korbak (currently at Anthropic)

Track record (project leader - Piotr): attended University of Warsaw's CS program myself + my high school classmate is currently teaching there as a PhD. Created a profitable 2-person company at age 20. Recruited 300 people in 2 months for a non-profit project via unpaid Facebook marketing in early 2020. Worked 5+ years as a software dev, briefly as a junior ML dev, attended ML4Good, been familiar with the EA movement & AI alignment ideas since 2014.

What are the most likely causes and outcomes if this project fails?

There is really no possibility for a complete failure. In the worst case I think we would manage to convince much fewer people than we initially hoped. But we plan to reach everyone we can, one way or the other (one could DM every student on Messenger or email if all else fails).

The most likely obstacles I can foresee would be someone from the faculty actively trying to prevent us from reaching the students. In that case, we would try to speak to the students directly, e.g. online. We would also try our luck with other programs and other universities in the city.

How much money have you raised in the last 12 months, and from where?

Zero. It would be my first grant. I provide my own salary, so one could say I have raised that. I'm able to pay myself because of my past investments (I had invested in real estate with the money earned as a programmer).

My income is only ~650$ net a month (which has purchasing power comparable to ~1300$ in the US), so I am unable to fund more than just my salary.

I like libertarian / randian aesthetics, am quite frugal and will not waste your money.

I just gave Piotr a small top-up for the purposes of running 2 hackathons (or events in the same spirit) in areas of AI Safety, eg making a satellite event for an Apart Hackathon.

I don't know Piotr much, and so can't be confident he'll execute well on this, but this seems pretty positive EV to me. This is very cheap, and I've met a surprising amount of University of Warsaw graduates in elite STEM circles (eg maths olympiads, Cambridge university, Jane Street, etc), so getting any of them excited about AI safety seems great to me. I think hackathons are a great way to energise and ground a group into actually doing practical things, rather than just reading or philosophising, and to help build momentum.