ControlAI’s Creator Outreach

How informing millions about extinction risk from superintelligent AI works

OpenAI has spent “low hundreds of millions” to buy a podcast to “accelerate global conversations around AI”. AI companies have embedded themselves in traditional media through corporate deals, and are funding super PACs to the tune of hundreds of millions to target politicians who support any AI regulation.

This is a propaganda operation, designed to improve the AI companies’ reputation, downplay the risks of the technology they’re trying to build, and avoid regulation.

On the other side, at ControlAI, we are proud to transparently work with creators to inform millions of people about the extinction risk from the development of superintelligent AI (ASI) and the necessity of an international ban on ASI. Through these partnerships, we inform the public about the current situation, our work to prevent the risks, and how to get involved and make a difference.

Recently Sabine Hossenfelder made a video on AI safety nonprofits’ creator collaborations, including ControlAI’s. The tweet about the video went viral, and as always on the internet, a slew of people reacted to the video title rather than its content.

Thanks to this wave of attention on creator partnerships, now’s a great chance to talk about how we work with creators.

Most people don’t know how the creator industry works, so we’ll start with a short overview, particularly about YouTube, followed by our standard process to work with creators. Afterwards, we’ll discuss this particular case with Sabine Hossenfelder that has gone viral.

How the creator industry works

Most major creators are professionals who make a living via the content they make. Income comes from a few sources including YouTube AdSense, Affiliate Marketing, Merch Stores, and Paid Subscriptions (like Patreon), but for many creators the majority of this income comes from Brand Partnerships.

If you’ve watched YouTube videos, you’re familiar with the transitions to e.g. VPN ad reads by the video host. For professional full-time YouTubers, a sponsorship on every video is usually a necessary part of their business.

The process of setting up sponsorships is a standard business partnership, requiring outreach from either the creator or the brand to the other, coordination on deliverables, price and contract negotiation, and project management. This usually means the time from initial outreach to video published can be from 30 to 90 days, or even longer depending on the project.

For many, especially smaller creators, this is quite a lot to handle on top of making the videos, so they hire managers to serve as the intermediaries, who then take a cut of the sponsorship deals. Most of the time when you reach out to a professional creator, a manager is the person picking up on the other side.

YouTube’s ecosystem is by far the most professionalized of any of the media platforms, though it still has its share of shady dealings. Short-form platforms like TikTok and Instagram have a far more rampant problem with undisclosed sponsorships.

Though in most cases disclosure is legally required for commercial advertising, there is a lot of gray area online, and very little enforcement and few consequences for those who bend the rules.

ControlAI has never engaged in undisclosed paid sponsorships with content creators. Every single video we sponsor mentions our involvement within the video, and in the description. We are proud of the work we do with creators and want viewers to know about it and to take action with our tools!

ControlAI’s Work

In case you aren’t familiar already, most of ControlAI’s work is focused on AI policy where we brief lawmakers about our campaign to ban the development of superintelligent AI to prevent the extinction risk it poses to humanity. In just about 2 years:

• We have directly briefed nearly 400 lawmakers across the US, UK, Canada, & Germany
• Our campaigns in the UK and Canada are supported by 170+ lawmakers
• We testified multiple times in the Canadian Parliament
• Our UK Artificial Superintelligence Security Bill was the first bill banning superintelligence development to be introduced in any legislature in the world
• We consulted on the first US bill to ban superintelligence development, announced in September 2026

The core of ControlAI’s ethos and strategy is the Direct Institutional Plan (DIP), which we developed. This plan is straightforwardly designed to prevent AI extinction risk by informing as many lawmakers, journalists, creators, and other influential figures as possible, as well as the public directly about the risks from ASI and the need to ban its development, and help them act. By design, this plan can be implemented by anyone or any organization to contribute to preventing human extinction.

Public awareness about AI risks lags significantly behind where it should be as a result of how quickly things are moving, and we believe that everyone has a right to know and understand what is happening in AI and the risks that are being forced upon them by those trying to build ASI.

ControlAI is all about action – by building common knowledge about the threat from ASI and a broad movement to ban it, we can make a difference. As such we do not support “doomerism”, where people discuss extinction risk from superintelligence as if it’s inevitable without providing any solutions. For audiences, we built world-class tools to contact your representatives across four countries. These messages can make a significant difference by showing lawmakers their constituents care about this issue.

In March 2025, ControlAI began conducting creator outreach as part of this strategy to inform creators about the issue and see how we could better raise awareness about AI extinction risk among the public.

Some creators are in a particularly fortunate position where they already have consistent sponsors that are happy for them to talk about anything they want, low production costs, or independent resources to fund their videos. In these cases we assist them with research, scripting, fact-checking, and connecting with high-quality interview guests.

This leads to some fantastic work with amazing creators like Professor Dave Explains, but not every creator can simply cover any topic they are interested in for free.

We’re super grateful for these creators, but of course we cannot expect everyone to work for free. Especially when they have high production costs (e.g., animations), or want to cover the issue but do not know in advance whether this topic they care about will perform as well or better than their usual content.

Incidentally, our work with creators has regularly led to outlier videos such as our sponsorships with Rational Animations and SciShow. A large part of this success is that we’re experts at explaining extinction risk from ASI, as this is the single issue our organization is aiming to solve. We know which explanations, examples, and analogies work, and which ones flop.

Given all this, we sponsor some creators, which in practice means buying a 60- to 90-second ad slot in their video and agreeing on the topic in advance so that backing the production makes sense for us.

In total, we’ve sponsored 9 paid videos and helped produce 9 unpaid videos all covering ASI and extinction risk with over 6.3M+ views combined. And beyond informing millions about these risks and how they can help avert them, over 3,000 people watching those videos have contacted their lawmakers using our tools (our main CTA).

As with our policy and media processes, the first step to working with many people is many more cold emails, and many meetings! So far we’ve contacted around 1,000 YouTubers across the space about their interest in covering this topic. From that outreach, we’ve briefed over 60 creators directly on this issue, with a combined subscriber count of over 110M.

You might be wondering, how do we find our partners, and what exactly do we ask of them?

Working with creators who actually care

ControlAI is an advocacy organization that focuses on a single issue, preventing the extinction risk posed by superintelligence. Like most nonprofit advocacy orgs that work with creators, we only want to work with creators who actually care about this topic. To ensure this, we always ask for, and usually require, direct briefings with the creator before going ahead with any partnership. We’re very upfront and tell them our goal isn’t to pay creators to parrot our talking points for cash: it’s to find and help those who actually care about the issue and produce good, informative content on ASI extinction risk.

ASI extinction risk is a complex issue that most people don’t know much about, including many creators. Our briefings offer the opportunity for creators to ask questions, discuss disagreements, and generally just learn more about what’s going on. We’ve even had a good few briefings where the creators are upfront that they probably won’t make a video and just want to learn more about the topic, which we’re happy to provide.

We solely advocate for a ban on ASI and not for generic anti-AI sentiment, nor certain AI policies, such as a moratorium on datacenters, that would undermine the economy and security of an individual country without directly preventing ASI. Creators we talk with are of course welcome to discuss their own views for or against any other AI topics, but if they aren’t on board with our campaign, we naturally don’t want to sponsor them for an off-topic video.

Put differently – we’re not in the business of sponsoring generic pro or anti-AI content for its own sake. If a creator wants to make a video expressing a broad range of anti-AI opinions, with a section on extinction risk from ASI, we’re happy to work together. If a different creator wants to talk about their AI coding project but also explain that ASI needs to be banned, we’re happy to work together.

Some creators simply don’t want to talk about ASI and extinction risk. A briefing with them directly makes the disagreements clear, and we do not partner with these creators.

Some people claim there is a strong incentive for creators to pretend to agree with us for the sake of making a partnership work for monetary gain. But given that we’re a nonprofit and pay less than a standard corporate sponsor, this tends not to be an issue. If a creator disagrees on the topic, it would likely cause friction with their audience, and they can simply opt for more lucrative generic brand partnerships selling a product.

So, once we find someone who supports our campaign and wants to make a sponsored video, what do we ask of them?

Our Partnerships

We like to keep things as straightforward as possible with our partnerships. Once the creator is on board, we first settle on a video topic via a simple pitch that can tie into ASI extinction risk. Then we sign a sponsorship contract. Here’s what that looks like:

The sponsored segment transitions from the content of the video about the problem to what we can do about it, going over ControlAI’s campaign and calling on the audience to contact their lawmakers using our tools at controlai.org/take-action. As is standard with YouTube sponsorships across the board, we get editorial input for these clearly marked sections of the video.

We usually require one round of non-binding script review, where we provide helpful feedback to the creators to improve the video script based on our extensive experience with videos on this topic. Creators are free to take or leave any or all of our feedback, as we’re already committed via the contract and don’t have editorial control. We might also have a similarly structured round of full video review for a second round of feedback.

Naturally, we do not want in any way to associate with and sponsor videos that call for violence, that contain factual errors or misleading claims, or that do not cover the issue we are focused on, and script review is a crucial step for us to ensure this. If we decide to pull out at this point, we usually pay the creator a cancellation fee, depending on the amount of work the creator has already undertaken.

Creators maintain full editorial control over the main content, with two key exceptions:

First, we require that they do not endorse any AI companies in the video. At ControlAI we have always been very clear that the threat from ASI comes from its development, no matter which actor is doing it: whether Anthropic, OpenAI, or any other company or entity. These trillion-dollar companies already have no trouble paying for their own marketing budget themselves, and we will not subsidize it.

Second, we require that they cite either the Statement on AI Extinction Risk or the Statement on Superintelligence at any point in the video to showcase the widespread expert warnings on this issue. This serves as a simple bar to ensure that they are actually covering AI extinction risk, rather than generic anti-AI sentiment, or other AI issues, which our organization does not focus on. It’s an obvious and easy citation to make if you are actually engaging with the issue, and an awkward one to make if you aren’t.

The partnership is of course disclosed in the sponsored segment or even earlier depending on the video, and the top line of the description and pinned comment discloses it alongside the CTA link.

Not much more to it, that’s the whole process!

Our Engagement With Sabine Hossenfelder

Normally we would not publicly discuss our interactions with creators. In this case, given the creator has already chosen to give her side of the story, and named us publicly, it makes sense to share ours as well.

This is also an informative case study into why we prioritize meeting with creators directly and why we’ve structured our approach as we have.

We reached out to Sabine Hossenfelder in July along with hundreds of other creators as part of a new outreach campaign. Her manager replied, and was the only person on her side that we ever spoke to. He told us Sabine supported our campaign and wanted to collaborate, but she would not meet with us.

We emphasized the importance of meeting directly with creators and insisted on wanting to meet Sabine, to which the manager assured us that Sabine was already bought into the issue and that was why she wanted to collaborate. Given these assurances and her well-informed past coverage of the issue, we decided to experiment with making an exception to our direct briefing requirement, and agreed to move forward.

We sent over our standard partnership requirements, to which no concerns were ever raised. The manager replied with a full script for a video that was off topic for us, that also endorsed an AI company, and contained no citation of either the Statement on AI Risk or the Statement on Superintelligence. This made clear to us that something had been lost from communicating solely through the manager, and that what we are looking for in a partnership had not gotten through.

We replied that partnering on this video didn’t make sense for us and asked whether Sabine had any other videos in the works that might more closely align with our campaign, such as one covering the then-fresh news of the OpenAI-Hugging Face attack. The manager replied that the angle was “unlikely to resonate well with her audience or fit naturally with her content” and that we couldn’t “request any changes”.

Then the next day Sabine posted a video opposing campaigns that would slow down AI development. It’s fine for Sabine to have a different perspective, but this was not the impression we had been given by the manager, and all of this would have quickly been revealed with a direct meeting.

Given Sabine’s position in the new video and off-topic script proposal, the partnership clearly was not the right fit for us, and so we politely declined, but left the door open for if Sabine’s interests changed. The manager did not respond.

Nevertheless, three weeks later the manager reached out to ask if we’d be interested in sponsoring a video about the OpenAI-Hugging Face attack, the very topic we proposed which previously was “unlikely to resonate well with her audience or fit naturally with her content.”

The next day, before we replied to the manager’s email, Sabine posted the video about our interaction that kicked all this off. Among other things, she said: “they wanted to tell me what the video was supposed to be about and what I was supposed to say down to the sentences I should read and references I should quote. I said no.”

First, we never got to speak with Sabine herself. Second, given how the manager misrepresented Sabine’s position to us, we assume he likewise misrepresented our position to Sabine. Once again, our only requirements were the standard ones described above, not full editorial control over everything she said and referenced.

Interestingly, the manager followed up with us again this morning (Sep 8th) to see if we still wanted to sponsor the video (which appears to have just been published). This is very strange, as usually when people disagree and make videos about turning down partnerships, they don’t then continue to ask the same organization for partnerships.

It seems unlikely that Sabine herself would even agree to work with us now. Our best guess is that either the manager is not paying attention to Sabine’s videos, and/or is trying to force a deal through even if it doesn’t make sense for either of the parties.

Regardless, this experience serves to thoroughly reinforce our rule that we always meet directly with creators before partnering with them. And despite all this, we would still happily meet with Sabine directly to discuss the issue if she were interested.

Conclusion

ControlAI is working to raise public awareness about extinction risk from superintelligent AI. We do this alongside creators who are also concerned about this issue and want to help. We’re happy to sponsor creators’ videos covering this topic and give their audiences a CTA to contact their lawmakers.

If you’re a creator who’d like to learn more about or cover this issue, we’d love for you to reach out to maxwinga@controlai.org, to see how we can work together to inform millions more about this urgent issue that impacts everyone, and help them take civic action to ensure humanity stays in control.