Hi, im Percy. Im a software engineer from Portland, Oregon.

As many of you, i am concerned at the pace of advancements in the sphere of artificial intelligence. The release of GPT4, and article's by Eliezer Yudkowsky, has become my wakeup call to take action. To slow down the speed of ai development.

In approximately 2 weeks, i am planning to fly to San Francisco and organize a week long picket around OpenAI's headquarters.

My goal would be to get attention of the news, media, journalists, get positive coverage, and inform more people about the real dangers associated with AI, and ultimately recruit more supporters to the cause.

Right now i am in the planning stage, on how exactly to approach this. My current ideas are:

Contact and inform journalists beforehand, so that they will be there in time to give me coverage.

Develop clear message, and demands, and best approach to this protest. Clear explanation of ai dangers that anybody can understand.

Make props: signs, papers, slogans, etc, that catches the eye.

Improve and practice my public speaking, get feedback.

Get help, ideas, feedback from the people and communities that share my worries. From individuals, members of lesswrong in the Bay Area. Or anybody. Be it financial, a place to stay, wisdom, connections, help from people who work in PR, feedback, etc.

We are living in extraordinary times. We could be the last generation of humans to ever live. We owe to our loved ones, to our children, to the entirety of humanity, a duty to avert a crisis, and ensure a brighter future.

So, who is with me?

Please, give me all the help you can give. Assist me in my act.

New Comment
15 comments, sorted by Click to highlight new comments since:

I don't think this is a good idea. You and others may reasonably disagree, but here's my thinking:

Protests create an us vs. them mentality. Two groups are pitted against each other, with the protestors typically cast in the role of victims who are demanding to be heard.

I don't see this achieving the ends we need. If people push OpenAI to be for or against AI development, they are going to be for development. A protest, as I see it, risks making them dig in to a position and be less open to cooperating on safety efforts.

I'd rather see continued behind the scenes work to get them to be more cautious, e.g. like the work ARC Evals is doing. It seems more likely we can have a positive influence by working with them rather than directly opposing them.

Protest seem more effective when it's an expression of a mass movement. It's not clear to me there's a mass movement to oppose what OpenAI is doing, so it's hard for me to see what positive impact a protest would have. Seems much more likely to make things worse than better given the conditions under which you would be protesting.

Protests create an us vs. them mentality. Two groups are pitted against each other, with the protestors typically cast in the role of victims who are demanding to be heard.

Dilemma ("choose a side") is a principle of non-violent direct action; why is an us vs. them mentality necessarily a bad thing? Do you oppose protest in principle?

If people push OpenAI to be for or against AI development, they are going to be for development. A protest, as I see it, risks making them dig in to a position and be less open to cooperating on safety efforts.

Would you say this about the climate movement pressuring fossil fuel companies to transition away from fossil fuels?

I'd rather see continued behind the scenes work to get them to be more cautious, e.g. like the work ARC Evals is doing. It seems more likely we can have a positive influence by working with them rather than directly opposing them.

I think we need both – here's evidence for the radical flank effect.

It's not clear to me there's a mass movement to oppose what OpenAI is doing, so it's hard for me to see what positive impact a protest would have.

Strongly agree – and this is how all mass movements start, no?

These are some of the reasons I don't want to join your protest: You've done basically zero of the hard work required to rally people behind a successful protest (other than write this announcement). You'd really need a concrete policy ask before I would think of joining your protest, and show that you're going to be able to coordinate a bunch of work from lots of people. More generally I don't really like the dynamic where the first person to say "me" is suddenly able to direct a bunch of free-energy, even if they probably aren't able to follow-through. Also nothing in your post indicates that this would be very rationalist or LW in nature.

That said I like that you want to do something to prevent a potential extinction-level risk. I do think you could do something smarter than this attempt, and try harder to figure out what might work.

You've done basically zero of the hard work required to rally people behind a successful protest (other than write this announcement)

Isn't this how most social movements start – with a single protest, attended by a small number of people?

You'd really need a concrete policy ask before I would think of joining your protest

I think this is why Percy posted here: to discuss what that might look like! And perhaps he doesn't need specific demands – look at Occupy Wall Street as an example of a movement with underspecified/vague demands that was effective in some ways (and failed in others).

More generally I don't really like the dynamic where the first person to say "me" is suddenly able to direct a bunch of free-energy

Again – surely this is how all social movements start? This picket won't be perfect; in my view it will highly likely be better than nothing.

I do think you could do something smarter than this attempt, and try harder to figure out what might work.

Do you have any suggestions?

Hi Alistair! You might want to look into more strategic ways of planning activism work. It's true that many social movements start becoming visible with protests, but there is a lot of background work involved in a protest, or any activism.

It looks like your goal is to slow down AI development.

First, you'll want a small working group who can help you develop your message, analyses, and tactics. A few of your colleagues who are deeply concerned about AI risk would work. When planning most things, it's helpful to have people who can temper your impulses and give you more ideas.

I see that you want to "Develop clear message, and demands, and best approach to this protest. Clear explanation of ai dangers that anybody can understand." I recommend doing this more than 2 weeks out from launching a campaign, with help from your working group. There are many important talking points you can use around AI risk, but if you just pick one clear phrase for your campaign, it can get more traction.

After you're clear on the one most important message for you to spread right now, you want to know who you need to tell and who can help you tell it. This is the time for a stakeholder analysis. Be clear on:

  • Constituencies (who you represent the interests of)
  • Allies
  • Opponents
  • Targets (who can change things)
  • Secondary targets (who can influence them)

Then, and only then, you want to think of which tactic is best for you to influence your target towards your goal. A protest might not be the best way to convince them, for many reasons that I'll leave up to people who know more about AI and the stakeholders. Protests are one tactic, but so is a well-planned email campaign to officials (with a template, gone through rounds of revision and feedback, for your participants), a social media campaign with text and images about the risk, or visual/literary art about a hypothetical future where AI development does not slow down.

After you've selected a tactic (and made a plan for the project that has been reviewed by people who are experienced in AI safety), you can organize your community to help you carry out that tactic as massively as you can.

This process from start to finish might take a few months, but it is worth getting it right the first time. Then, you will have less issues to fix, and you can build on your momentum and scale up. The more people who help you think through how to effectively influence your targets towards your goal, the better. It is best if you have some people to work with who are deeply familiar with your target. Networks are everything in organizing. Good luck.

Edit: I figured you might want me to tell you why I'm recommending all these other steps. I'm doing that because I'm seeing you receive feedback (from people more involved in the issue than I am) that this could cause harm. I saw you say "in my view it will highly likely be better than nothing" above. It might be worse than nothing. Hence the planning. It seems you want to act fast because this is an urgent threat, and I get that. But acting fast and making things worse is worse than planning for a few months and making things much better.

It seems to me like you would like to be able to succeed at preventing an extinction-level threat without having to be competent at anything. I think reality has higher standards than that.

Occupy Wall Street is exactly the wrong example to mention, and further makes me think you have no sense of what a successful historical protest looks like.

I'm not into random people showing up and trying to command the political force of a web forum that they've made zero contributions to and trying to direct it into a poorly thought out and haphazardly-aimed effort.

I am well-aware of the stakes here, but that doesn't mean bad plans suddenly work.

[-]RubyModerator Comment105

I'm the moderator who approved this first post. Like others, I don't think as stated this is plan is being conducted in a good way. However, I thought it might be good for Percy to get feedback and pushback. There's been less explicit feedback, but the -30 karma is in indication that people in general do not think this is a great plan.

Percy, you've gone and made another post continuing with your plan and disregarding the response from this one (and being more concrete without having done the work you say in this post needs doing). I've drafted that post as I don't want new users to be able to use the platform to make and coordinate plans that don't seem very good especially if they're not engaging with the feedback they're getting.

If you like, you can still engage in the comments here if you want to defend your plan. I urge you to be wary unilaterlist's curses here.

A little more compassionately, it's pretty horrific that some humans seems just totally willing to engage in activity that could destroy everything for everyone, and wanting to counter that seems very very understable. I recommend my A Quick Guide to Confronting Doom and say the need do something doesn't mean this thing needs to be done.

This seems to be another one of those instances where I wish there was a dual-voting system to posts. I would've liked to strong disagree with the contents of the post without discouraging well-intentioned people from posting here. 

a week long picket around OpenAI's headquarters.

This is an unexpectedly creative way to screw this up. Planning a protest to be a week long means that most would-be attendees don't know when they should be there, and will show up at a random time in a week-long interval, see that there's no one else there, and leave.

If you want this to be at all successful, you need to pick a specific date and time. It's fine if you're there more often than just then, but please, for crying out loud, don't position yourself as an organizer and then create ambiguity about basic logistical details.

A somewhat messy braindump[1] of my thoughts on this style of approach:

I'm excited to see you interested in participating in the distributed network of humans trying to ensure AI goes well! I don't currently think that protests are a useful collective action; they barely work to petition, and I suspect are heavily amplified by toxoplasmosis of conflict. Synchronized, organized expression of demand works best when a large group of people are a dependency of the system, because a protest shows that those engaging in the collective action of showing up could also engage in other collective actions. But openai doesn't need you to like them, and in fact the power imbalance with openai is already frighteningly intense. Reaching the sort of many-vs-few bargaining position that collective action seeks to create seems to me to mostly amplify thoughtless mammal-conflict response in the openai team. Strategic collective action probably looks more like figuring out ways to get many many programmers to think about how to survive the superintelligence transition at a technical level. If you're interested in that form of capability-selected collective action, I'd love to talk about it, and would be enthusiastic to upvote a post discussing it. I'd also be interested in discussing what sorts of interpolitical societal discussion[2] are valuable. I'm a big fan of solidarity as a base on which to build, and though I disagree with the global left in some ways about how to best achieve solidarity, what part of the global left doesn't? in particular, though, in this situation, I think it is important to impress upon those used to being in charge at the top that all-of-humanity solidarity, not just human class solidarity, is needed to prevent the rise of catastrophic AI. If you'd like to do political organizing for the task of preventing the rise of dangerous AI, I'd suggest reading up on how ai policy people are currently thinking about it. It's unusually important to avoid polarizing because right now both left and right are terrified of AI for basically the same reason, though of course they're phrasing it differently ("preserving our values and preventing extinction" on the right, as usual, vs "preserving worker rights and autonomy" on the left, again as usual; the unusual thing is that this time, these refer to the same thing!)

  1. ^

    I was one of the main examples criticized in "No, you need to write clearer", oh well i guess,

  2. ^

    eg, memetics, but that dismisses how much people can think about things before contributing