Creator partnerships for impact.
Expert briefings, events and content partnerships that give trusted creators what they need to help millions understand advanced AI and its effects.
About Cascade
Cascade helps AI safety organizations work with creators to reach people beyond their usual channels. We start with the audience you want to reach and what you want them to understand or do, then decide together whether creators are the right fit.
We coordinate paid partnerships, expert briefings and events—from finding creators and shaping the work to factual review, disclosure and measurement. By maintaining these relationships and processes across organizations, we aim to make each project easier to start and strengthen the quality and trust behind the next.
The problem
Some organizations already have strong communications teams. Others need creators for an occasional campaign, research release or urgent story, but building the relationships and processes for a handful of projects is costly. Cascade maintains that capacity across organizations, so each partner can use it when there is a clear need.
Editorial independence
Partners define the objective, factual boundaries and agreed deliverables. Creators decide how to tell the story and retain their own opinions. We provide sourced briefings and expert access, with factual review of paid work focused on accuracy and agreed requirements—not scripting a creator’s conclusions.
Transparency
Audiences should be able to see who funded a piece of content and that the creator was paid. We build clear disclosure into every paid partnership, with fees, deliverables and responsibilities agreed upfront. Taking part in a briefing does not imply endorsement or an obligation to post.
Audience fit
We start with the people a project needs to reach. Creators are selected for their audience, interests, credibility and ability to handle the subject—not follower count alone. Vetting and suitability checks help partners assess the fit before committing to a campaign.
Understanding and action
Reach is only part of the picture. We agree the intended outcome and success measures before a campaign begins, then look at what people understand or do next, where we can measure it. Results are reported against those goals, with limitations made clear and lessons carried into future work.
How Cascade works
We start with who you want to reach and what you want to achieve. Then we match you with creators and shape the work together, coordinating introductions, expert briefings, factual support and delivery. We agree how to measure results before the campaign begins. Creators keep their voice, and paid partnerships are clearly disclosed.
Research
Researchers explain AI risks and ways to address them.
Partnerships
Creators bring the evidence into stories their audiences care about.
Understanding
People connect AI risks to their lives and communities.
Action
Audiences find ways to learn more and participate.
Informed decisions
Wider understanding can help shape decisions about AI.
- Creators in our network
- Campaign types to choose from
We’re building lasting relationships between creators and researchers to help audiences understand AI risks and find ways to act.
Work with us
We are looking for a small number of AI safety organizations with a research result, campaign, event, public education goal, or audience they want to reach. We are especially interested in projects where the organization can define the intended audience, desired action, and success metric in advance.
If you have a project where creator partnerships might help but do not want to build the sourcing, relationships, contracting, briefing, production, and measurement infrastructure internally, we can run that process with you. For time-sensitive work, we are able to launch a rapid-response campaign within one to two weeks of an incident.
Cascade is newly formed and operates under a fiscal sponsor’s exemption until we spin out as an independent 501(c)(3). During this initial phase, partner organizations pay creators directly and Cascade’s coordination is grant-funded, so there is no fee to 501(c)(3) partners.
Our immediate goal is to run a small number of partnerships and learn where maintaining creator relationships and campaign capacity across organizations provides the most value.
The team
Caitlin Yardley
Co-Founder & Director of Talent and Communications
Journalist turned AI-safety creator (@futurebriefing). Co-organized BlueDot and Lighthaven creator events. Existing network of creators from PDKU, the Frame fellowship, and adjacent networks. Creator work here and here.
Sofia Mendez
Co-Founder & Director of Strategy and Partnerships
Leads AI strategy at LLPA, working on AI/AGI preparedness across a global network of education orgs in 62 countries spanning government, SMEs, and universities. Previously a software engineer working on production systems serving 1M+ users. Creator work here and here.
Frequently asked questions
How do you balance creator independence with factual accuracy?
Our goal is not to give creators scripts and make sure they strictly follow them.
We identify creators whose existing content and audience make sense for the overarching goal of the project, then give them enough context, sourcing, and access to experts to communicate the issue accurately in their own voice.
There is a real tradeoff here. If content is optimized only for reach, complicated arguments can become simplified or distorted. If it is too dense, creators may not be able to make it engaging enough to reach mainstream audiences in the first place: research on social media engagement finds that easier-to-read content tends to receive more engagement.
Our approach is to preserve the substance of the argument while making it understandable and engaging to a mainstream audience. Creators receive sourced material and expert access. Every paid deliverable gets a factual review before publication, and that review is limited to accuracy, sourcing, disclosure, and other requirements agreed with the partner organization. Content produced after briefings or events, where no payment is involved, is not reviewed.
What if a creator says something we disagree with?
You approve every creator before a campaign begins, and the requirements you set at the start (factual boundaries, deliverables, disclosure) are written into the creator’s agreement.
If a paid deliverable doesn’t meet those requirements, it is corrected before publication. If the disagreement is about the creator’s opinion or framing rather than facts or agreed requirements, we won’t ask them to change it. You are never obligated to reshare or promote a piece of content, and you can choose not to work with that creator again.
What if a creator becomes controversial after we’ve worked with them?
We plan to prevent this through rigorous creator vetting, which has become industry standard for commercial brand deals after brands faced reputational damage from poorly vetted partnerships. We will also conduct quality assurance before any video goes out. In addition to this, we are already seeking a public relations advisor and, with future funding, would bring dedicated PR capacity in-house.
What kind of creators do you work with, and how will you make sure they understand AI safety well enough to produce accurate content?
We want to work both with existing AI safety creators and with mainstream creators who are interested in learning more.
For fast-turnaround work, we will generally prioritize AI safety creators or creators we already have a relationship with and can brief quickly. For each campaign, creators receive a sourced briefing, access to relevant experts, and factual support during production.
For rapid-response work, where a partner exists, the partner names a single reviewer for accuracy and sourcing before publication. For Independent Comms Fund work, Cascade fills that role and applies the same standard. Corrections follow the same process as any other campaign.
The goal is to maintain a pool of creators who already have enough context to respond quickly rather than starting from zero when an incident or narrow attention window occurs.
We are also interested in informing and supporting creators who do not normally make AI safety content, or whose audiences do not normally encounter AI risk discussions through usual AI safety channels, and bringing these ideas and ways to engage to audiences that AI safety communications do not currently reach.
Can you work with embargoed or confidential research?
Yes. Creators briefed ahead of a release agree to confidentiality terms, and publication is timed to your announcement. Nothing is published before the embargo lifts.
What does transparency look like in practice?
Every paid campaign should make it easy for someone watching the content to understand who funded it, why the creator is participating, and that the partnership is paid.
There is already strong demand for this, particularly around paid advocacy. In a 2026 study of 906 U.S. social media users, 88% agreed or strongly agreed that creators should disclose when they receive payment from a political organization, party, or campaign.
Commercial brands faced similar problems when sponsorships were not clearly disclosed. In a survey of more than 3,720 U.S. consumers, 70% said they felt negatively about or deceived by an undisclosed brand partnership.
There is also little evidence that clear disclosure has to come at the cost of performance. A 2026 analysis of 583,000 sponsored Instagram posts found no meaningful engagement penalty from explicit #ad disclosure.
In practice, that means platform paid-partnership labels where relevant, clear disclosure in the caption, and, where appropriate, disclosure in the video itself, following established disclosure guidance.
What do we get at the end of a campaign?
A results report measured against the success metric we agreed at the start. Depending on the campaign, it covers reach, the number of people who took your call to action, conversion rate, cost per qualified action, how performance compares with your existing channels, a comment sentiment and topic modeling, and disclosure compliance across every deliverable.
Do you work with organizations that aren’t 501(c)(3)s?
Our current offer is designed for 501(c)(3) organizations. If you’re a different kind of organization, get in touch through cascadeagency.org and we can see what’s possible.
How to start
If you have a project in mind or are interested in learning more, fill out the short form at cascadeagency.org and we'll schedule a call to talk through your goals and how we could support them!