Hide table of contents

We have 4 winners: Have a look at aisafety.art !

Notes on the votes, regarding feedback from the community:

I used STAR voting to process the votes, as some people kindly pointed out, and I also made sure by grouping the votes of logo variations that designs with many variations are not disadvantaged.

EDIT: vote here: http://vote.aisafety.art from 1st July - 14th July '23 11pm UCT 
and see page aisafety.art

EDIT: received all submissions, timeframe for voting will be delayed by about 2 weeks as one of the organisers wanted to add  more designs from a different source to the voting pool, but this won't affect prize distribution. (19th June 2023)

EDIT: extended deadline is Saturday May 27th 2023 11:59 PM UTC

This contest is about creating a logo to represent AI safety research. 

I’m being supported by an independent funder and people from nonlinear.org and am taking care of administration and communication.

We’re doing an open process (voting) and want to involve different people through the contest to get a better logo design and to increase the chances of adoption.


1) rough concept ideas contest

deadline: May 15th 2023

We’re looking for different quick and great design ideas for this round. Not everything has to be spot on yet, this will be done in 2) and 3) .

Submit as many rough logo design ideas as you want as a .PNG and .SVG file to this Google Form: https://forms.gle/6fwVYbPBrqVdCHax6


2) feedback & voting on best design ideas chosen by jury


Via another Google Form, this post will be updated with the link once there’s a selection by the jury of 5 - 10 contest submissions available.

Only accounts with at least 100 Karma on forum.effectivealtruism.org or lesswrong.com are allowed to vote.

With a $ 1.5k prize pool for the best design ideas from the contest which are being used further, this might be one design or multiple as decided by the votes & the jury, with the prize money spread out evenly. 

We’re open for someone to increase the prize pool, please send me an email!


3) polishing of winning design(s)

We’ll hire a great designer for creating the final logo from the winning design(s).

After the completion of the contest the final logo will be publicly available for free use at www.aisafety.art .
If you have any questions please reach out to me at ai_logo@cipriani.studio 


I’d like to disclose here that I’m also trying to enter an idea myself, however I won’t be able to decide on my own which designs are considered for polishing.

Composition of the jury: two people from Nonlinear, an independent funder, me (Adrian Cipriani)





Sorted by Click to highlight new comments since: Today at 10:10 PM

Suggestion: use a well-designed voting system such as STAR voting, approval voting, or quadratic voting.

Thanks for suggesting @RomanHauksson , so far something along the lines of approval voting for multiple winners seems favourable to me.

The current form appears to only allow uploading image files; I can upload a PNG, but not an SVG. This is probably just as well in my case, as the SVG only makes it more painfully obvious that I have no idea how to use Inkscape, but it seems like unintended behavior you might want to change.

Changed it, thanks for noting. You can add it now :)

Really excited to see this happening!

I enjoy doing this kind of thing and would enter several entries if you removed the requirement for a .svg file. It creates a pretty significant amount of unnecessary hassle - I'd understand if it was for a finished product, but I'm not going to use vector for sketches.

Adrian - good idea; I hope you get some excellent submissions.

Are you open to logo designs by AI systems? It might be ironic if a logo designed with the help of Midjourney or Stable Diffusion ends up winning.

Thank you @Geoffrey Miller ! Yes, we expect some people will use Midjourney as a source for inspiration. However, we expect submissions to be vector graphic files (.SVG) as well.

I’ve tried designing logos on Midjourney before. I’ve personally found it to be off mark in all of the cases. It would be really cool if the logo that won did use AI though.

PSA: The form accepts a maximum of 10 files, that is, 5 design proposals maximum (because each proposal requires uploading both a .png and a .svg file).

I added another field in case anyone wants to submit more than 5 different ideas in one turn.

Also am I right in thinking that you're entering for prize money as well as making decisions on who gets the money? 

Yes, however my vote is just 25% as we're in a team of four people.

I think it's bad if you have any advantage at all over any other entrants. I imagine you're also in a good position to exert influence over the other jury members if you wanted to. All things considered I think it would be best, or at least the most fair, if you were not allowed to financially gain from this. 

I would appreciate others' input on this as well (eg via agreement up/downvote).

Hello @AnnieAB thanks for bringing it up, I didn't submit any ideas in the end. Hope this helps!

What's the exact deadline time tomorrow, please? 

We decided to extend the deadline until Saturday 27.5.23 11:59 PM UTC because we received only a little more than 20 different logo ideas so far. Hope this helps!

Aw, I stayed up late for nothing 😅 

You might need to spread this more widely than the forum if you haven't already.

Only accounts with at least 100 Karma on forum.effectivealtruism.org or lesswrong.com are allowed to vote.

I'm a little confused by this. What's the motivation for using a karma threshold to decide who does and doesn't get to vote?

Hi Quinn,
this seems like a hurdle to prevent the use of sock puppets to prevent mass voting on your own logo idea. It is common that online voting gets attention of trolls, and they vote for something funny or extreme instead of good. This would partly also be the reason they use voting + the jury.
Voting gets the community engaged, they want to help their favourite creator and can also be part of the jury themselves.

Note: I am not a part of the organizer team and had no influence in the decision to use a threshold or on its height.

Yes, that's basically our motivation. Thanks @Felix Wolf for explaining!

@Felix Wolf Thanks for taking the time to explain, Felix. This makes sense now.