On November 24, 2025, iHeartMedia began appending a phrase to the top of every hour across its 850+ radio stations. The phrase was "Guaranteed Human." It appeared in the legal ID—the contractual announcement that identifies the station to the FCC and to its audience—as a brand promise, not a disclaimer. Tom Poleman, Chief Programming Officer and President of National Programming Platforms, explained the reasoning in a memo to staff: "We don't use AI-generated personalities. We don't play AI music that features synthetic vocalists pretending to be human. And the podcasts we publish are also Guaranteed Human."
The statement was unremarkable in its content and extraordinary in its necessity. iHeartMedia, like every major broadcast group, had access to synthetic voice technology. It had been approached by vendors. It had evaluated the economics. What it concluded—publicly, at scale—was that the efficiency of synthetic voice was worth less than the trust of a real one.
Lisa Coffey, iHeart's Chief Business Officer, cited internal research at CES 2026: 70% of consumers use AI tools, yet 90% prefer media from real humans. Ninety-two percent say nothing can replace human connection, up from 76% in 2016. Ninety-six percent found "Guaranteed Human" content appealing in testing. Nine in ten believe human trust can't be replicated by AI. These numbers are not surprising. They are, in the current moment, almost trivially predictable. What is notable is that iHeartMedia is the first major U.S. broadcast group to formally codify a no-AI-voices policy and market it as the primary brand attribute. The competitor Audacy went the opposite direction, partnering with ElevenLabs to build synthetic voice libraries for programming and ads, framed as augmentation. Both companies believe they are right. The market will eventually resolve which interpretation of the consumer is accurate.
In October 2025, Tega Brain—an Australian artist, environmental engineer, and NYU associate professor—released a browser extension called Slop Evader. The technical mechanism was simple: it appended a Google date filter to every search query, restricting results to content published before November 30, 2022—the public release date of ChatGPT. For platforms without native date filtering, it used Google to search those platforms by date. It was, Brain acknowledged, "the simplest, dumbest way to refuse it." It did not create a new search algorithm. It automated a function Google already provided.
"This sowing of mistrust in our relationship with media is a huge thing, a huge effect of this synthetic media moment we're in." — Tega Brain
Slop Evader is best understood not as a product but as a diagnostic. It reveals something about the current state of the internet that users felt necessary to construct a temporal firewall. The tool reached several hundred thousand installs. Its GitHub repository attracted contributors. Brain's own Instagram post about it received over 10,000 likes. The response was not fringe: it was mainstream enough to generate coverage in The Register, 404 Media, and WebProNews. The question implicit in every install is: what do users believe the post-ChatGPT internet has become?
The answer, according to Copyleaks data, is that AI-generated content on the web increased 8,362% between November 2022 and March 2024. Originality.ai's ongoing study of 500 popular keywords found AI content in Google top-20 results rose from 2.27% in February 2019 to 17.31% by September 2025—peaking at 19.56% in July 2025. Nineteen percent of all Google reviews written in 2024 were AI-generated, a 279% increase from 2019. The internet that Slop Evader's users are trying to return to was real, and it was finite, and it ended.
Pinterest's experience is instructive as a case study in what happens when a platform misjudges its own audience's tolerance for synthetic content at scale. Pinterest began integrating generative-AI pin creation tools in 2022 and, under CEO Bill Ready, pivoted toward an AI-powered visual-first shopping assistant model. By 2025, AI-enhanced ad placements were blending organic and promotional pins, with sponsored content reaching 31% of the feed. The platform reported 619 million monthly active users and revenue growth. The stock fell 13% after-hours in February 2026 because of weak ad-pricing outlook and doubts about the AI strategy.
The CNN Business investigation published November 22, 2025 documented the divergence: overall MAUs up 12% year-over-year, but dedicated core users leaving. Session length was down 12% year-over-year in Q3 2023. Organic content impressions fell to approximately 55% of the feed versus 78% in 2019. Ad pricing dropped 19% year-over-year in Q4 2025. Users interviewed by CNN explicitly blamed AI-generated content and algorithmic shifts for alienating them. Casey Fiesler, a researcher at the University of Colorado Boulder, told CNN: "They used to see a lot of human-created content that they found inspiring, and now it's just a lot of very non-human, perhaps not inspiring content." CEO Bill Ready acknowledged there is no way to catch 100% of AI-generated content. The platform optimized for what it could produce efficiently and discovered that its core audience was optimizing for something else.
Super Bowl LX, broadcast in February 2026, became, in retrospect, the moment the advertising industry received systematic feedback on its AI creative strategy in aggregate. The campaign-level data was already available: Coca-Cola's AI-generated holiday ad had been widely mocked. Toys "R" Us's Sora-produced ad, depicting founder Charles Lazarus as a boy, was described as soulless. H&M, Skechers, and Guess faced criticism for AI brand ambassadors. The Super Bowl compressed this feedback into a single evening.
Meltwater's analysis found approximately 52% of all Super Bowl LX ad sentiment was negative. About 50% of AI-related ad mentions were negative, describing the content as uninspired or AI slop. Dunkin's AI-de-aged 90s nostalgia campaign dominated AI conversation with 37% share of voice but received 42% negative sentiment. Svedka ran what may have been the first primarily AI-generated Super Bowl ad—a fembot robot theme—and achieved 63% of AI commercial engagement while generating largely neutral sentiment, which at this moment constitutes relative success. Artlist's five-day production of AI-generated animal clips was widely identified as the worst example. Meta's Ray-Ban Meta Glasses ad scored minus-15 on sentiment. The pattern was consistent: the more visibly AI the execution, the more negative the reception.
"I open up Pinterest and suddenly notice that half of my feed are these incredibly idealized faces of women that are not real." — Tega Brain
The consumer recalibration documented across these examples is real, measurable, and growing. The Billion Dollar Boy/Censuswide study conducted June-July 2025 found that consumers saying AI is negatively disrupting the creator economy rose from 18% in 2023 to 32% in 2025—an increase of 78% in two years. Consumer enthusiasm for AI creator output dropped from 60% to 26% over the same period. Gen Z preference for no AI in creative work reached 54%. Trust in influencer recommendations fell from 34% to 12% in one year. TikTok engagement rates dropped from 4.2% to 1.8%. Gen Z daily engagement with creator content fell from 67% to 28%.
These figures do not appear in any major advertising platform's metrics. They do not affect click-through rates. They do not appear in ROAS calculations. They are not visible in the dashboards that determine where $700 billion in annual digital ad spend gets allocated. They are visible in longitudinal sentiment analysis, in the slow migration of share-of-voice from synthetic channels to creator-economy channels, in the rise of Substack to 20 million monthly active subscribers and 17,000 paid creators, in the brands cutting influencer budgets by 60-80% and redirecting to owned channels.
Researchers at NYU and Wake Forest, in a paper published in the Journal of Business Research in January 2025, documented what they termed the AI-Authorship Effect: content generated by AI, when perceived as AI-authored, triggers perceived inauthenticity, which generates moral disgust, which reduces positive word-of-mouth and customer loyalty. Seven preregistered experiments found this pathway was stronger for emotional content than factual content. The effect was not trivial. It was not a rounding error. It was the mechanism by which sincerity became a competitive variable.
The Sincerity Gap is the name for what happens when an industry builds its production infrastructure around efficiency and discovers that a significant portion of its audience has been building its consumption preferences around trust. The gap is not rhetorical. It is not a trend piece. It shows up in engagement data, in sentiment trends, in the choices platforms make about what to label and what to disclaim, in the behavior of users who install browser extensions to opt out of the last two years of the internet. The people making budget decisions are not being told. The people being asked to evaluate the ROI of AI creative tools are not being given the dataset that would show what happens when audiences encounter those tools at scale. The Sincerity Gap is still occurring. The people who built the systems for measuring it have not yet built the systems for seeing it.