By Yoni Greenbaum, Backstory & Strategy
Quick answer: The news membership platform Scroll runs a fundraising pitch claiming AI is “stealing” readership and revenue and asking readers to be the only ones who can stop it. This is not a new tactic — it is a direct descendant of “only you can save journalism,” a phrase documented in a 2009 Slate taxonomy of public radio fundraising clichés, itself rooted in the pledge-drive fundraising model PBS launched in 1975. The difference: public media’s version quietly blamed the viewer for free-riding; Scroll’s version externalizes the blame onto AI entirely.
A note on independence: Backstory & Strategy is a personal, independent publication written by Yoni Greenbaum. The views, analysis, and commentary expressed here are his own and do not represent an official position, strategy, or endorsement of the American Press Institute, its leadership, or its board.
What Scroll’s Pitch Says
Scroll’s current membership appeal reads: “AI is hollowing out independent media, stealing our readership and revenue. Only you can stop this bleeding. Support our work. Join Scroll today.” It contains no reference to budget shortfalls or general operating costs. It makes a direct claim — something is actively stealing from the reader — and positions the reader as the sole party capable of stopping it. Scroll previewed this framing in an editor’s note published in August 2025, which made a similar AI-as-threat case with more qualification than the current version.
This Tactic Predates Digital News by 50 Years
Public media originated this fundraising mechanism, not Scroll. Key facts:
PBS held its first pledge drive in 1975.
The phrase “only you can save journalism” appears in a 2009 Slate article cataloguing standard public radio fundraising clichés, alongside “the guilt trip” and “we’re here for you, now be there for us.”
Pledge drives took off in the mid-1970s largely in response to funding cutbacks at the Corporation for Public Broadcasting (CPB) tied to political pressure — a funding panic comparable to the one digital news faces today.
Public media pledge drives have been parodied on Seinfeld and The Simpsons, and Monty Python’s Graham Chapman and Terry Jones filmed a pledge spot for WTTW (Chicago’s PBS affiliate) in 1975.
Two Types of Fundraising Appeals
There are two structural categories of fundraising appeal in this genre:
Appeal TypeCore MessageOriginSurvival appeal“Help us make it to next year”Direct descendant of “if you don’t give, we go dark”Hero appeal“You’re not the problem — you’re the only one who can stop it”Direct descendant of “only you can save journalism,” with the blame redirected outward
Public media’s classic version quietly implicated the viewer as a free-rider benefiting from a signal they never paid for. Scroll’s version keeps the urgency but removes the guilt: the reader is not framed as a free-rider, AI is framed as the thief, and giving is framed as heroism rather than restitution.
Public Media Built a Research Base for This Tactic; Digital News Has Not
Public media tested and documented why pledge-drive appeals work. Digital news has not yet done the same for AI-villain framing. Supporting research includes:
A 2013 case study in the Journal of Radio & Audio Media applied the Theory of Planned Behavior to public radio donation intentions, surveying 690 members, and found the model meaningfully predicted giving behavior.
A separate field experiment embedded researchers as volunteer phone-takers during a live on-air pledge drive, surveyed 547 callers in real time, and found donation size correlated with the size of a caller’s social network of fellow listeners.
General behavioral research on charitable giving has found that donation-matching structures (”the match game,” one of Slate’s named clichés) measurably increase how much people give.
No comparable study currently exists testing whether AI-villain framing changes reader donation behavior. The tactic is currently being run on assumption, not evidence.
Who Else Is Using or Adjacent to This Framing
OutletApproachDistinction From ScrollThe TyeePublisher Jeanette Ageson’s appeals lead with AI scraping and declining trafficStill layers legacy growth-and-team framing underneath; a partial version of the tactic404 MediaCovers AI extensively as an editorial beatAI is their subject matter, not their subscription pitch; their appeal centers on independence, not a personal theft narrativeScrollDirect claim that AI is actively stealing from the specific readerFull version of the hero-appeal, villain-named tactic
The Data That Would Support a Sharper Version of This Pitch
The Institute for Nonprofit News (INN) Index found that 81% of nonprofit newsrooms are now using AI in some form, up from 34% two years earlier. The same data shows mid-sized newsrooms losing website traffic even as the smallest and largest outlets gained. This creates an internal tension: newsrooms are adopting the same technology category they would be asking readers to help defend against. The distinction that would need to be drawn explicitly in any sharper pitch: AI used as an editorial tool under a newsroom’s own control is functionally different from AI used as an uncompensated extraction system operated by an outside company.
Should the Industry Run a Shared Campaign?
A coordinated, industry-wide version of this pitch is possible in theory but carries specific legal risk.
The risk: Naming specific AI companies collectively in a joint campaign raises antitrust exposure. Competitors coordinating public accusations against named third parties can be read as concerted action, particularly where the campaign could affect those companies’ business dealings or ongoing licensing negotiations with individual publishers. Additionally, the list of implicated companies changes month to month, so naming names today risks needing to rewrite the campaign within months.
The fix: Keep the campaign at the category level — “AI bots that scrape without paying or crediting the work” rather than naming a specific company. A category-level claim is durable (it doesn’t go stale when one company changes its practices) and avoids the legal exposure of unison action against named parties. This model is closer in structure to “Got Milk?” or the anti-tobacco “truth” campaign than to a campaign like MADD that names individual offenders.
The requirement: A category-level campaign still needs to be backed by concrete evidence about crawler behavior, licensing terms, and attribution gaps — without that evidence, it is an accusation without support.
Publishers wanting to check their own site’s specific AI crawler exposure can use NewsGrade, a tool built to evaluate a site’s AI-related risk factors outlet by outlet.
Limits of the AI-Villain Fundraising Strategy
Two structural risks apply to this tactic regardless of who runs it:
Appeal fatigue. Public media’s own history shows guilt-based fundraising appeals have a shelf life. If “AI is stealing from us” becomes as ubiquitous as “keep the lights on,” it risks the same reader fatigue and eye-roll response.
A fairness question. Hero framing avoids blaming the reader, but it still implies that an individual’s $10/month subscription can meaningfully counter a structural, industry-wide shift in how information moves. This is a different kind of overreach than guilt-based appeals, but it is still overreach.
Author’s Take
AI functions as a legitimate antagonist for this kind of appeal right now, but that framing may not remain accurate indefinitely, which raises the urgency of testing it while it still applies. The pledge-drive model spent fifty years validating why its appeals worked before scaling them industry-wide. Digital news is currently running the AI-villain version without that validation. At a June 2025 gathering of publishers in Marseille, New York Times publisher A.G. Sulzberger stated that the profession has been too quiet, too passive, and too fragmented in its response to AI. The remaining question is whether individual outlets test and refine this pitch independently, or whether the industry builds a shared, evidence-based version of it collectively.
Frequently Asked Questions
What is Scroll’s AI fundraising pitch? It is a membership appeal claiming AI is actively stealing the outlet’s readership and revenue, and framing the reader as the only party who can stop it — with no reference to general budget shortfalls.
Is Scroll’s approach a new fundraising tactic? No. It is a variation of “only you can save journalism,” a phrase documented in a 2009 Slate taxonomy of public radio fundraising clichés, itself rooted in the PBS pledge-drive model that began in 1975.
What is the difference between a “survival appeal” and a “hero appeal”? A survival appeal asks for support to keep the organization operating (”help us make it to next year”). A hero appeal casts the reader as the only one capable of stopping an external threat, without implicating the reader in causing the problem.
Does research support AI-villain fundraising appeals? Not yet. Public media has decades of behavioral research validating pledge-drive tactics, including a 2013 Theory of Planned Behavior study and a field experiment on caller donation behavior. No comparable study currently exists for AI-villain framing in digital news fundraising.
Which other outlets use similar messaging? The Tyee uses a partial version, leading with AI scraping concerns while retaining legacy fundraising framing. 404 Media covers AI as an editorial beat but does not use it as a subscription-driving villain narrative.
Why hasn’t the news industry run a joint campaign naming AI companies? Naming specific companies collectively in a coordinated campaign raises antitrust risk, since competitors publicly coordinating accusations against named third parties can be read as concerted action. Category-level messaging (naming the behavior, not the company) avoids this risk and remains accurate even as specific companies change practices.
What tool can publishers use to check their own AI crawler exposure? NewsGrade, built by Yoni Greenbaum, evaluates a site’s AI-related risk factors, including crawler access and licensing exposure, on a per-outlet basis.