The AI Deal Now Costs More Than It Pays: Write the Policy Before the Pitch

Business Insider has catalogued three ways creators are currently being punished over AI: taking a promotional deal from an AI company, being seen to have leaned on AI inside finished work, and being flagged as AI-generated by a platform detector when no AI was used ^[1]. For anyone whose revenue depends on public goodwill, that reclassifies AI from a tooling choice into a brand and talent exposure that needs a written policy attached.

The promotional case is the cleanest. Influencers who posted about attending a luxury OpenAI-hosted "Summer Camp" in early August drew comments from fans calling it "dystopian," "gross," and "morally bankrupt," according to Business Insider ^[2]. An OpenAI spokesperson told the publication the retreat was education-focused and that "creators are an important part of our community and how people get information and learn about our products" ^[3]. Separately, interior design creator Cliff Tan apologized to his audience after posting promotional content for the Chinese AI platform Dreamina, acknowledging the post had disappointed many of his viewers ^[4]. Eric Bogard, CEO of the talent firm UnderCurrent Management, told Business Insider there is "a PR issue with AI as it relates to taking people's jobs and environmental concerns around data centers," and that creators are hesitant to promote AI companies as a result ^[5]. Large AI companies have substantial budgets for influencer marketing ^[6], so the binding constraint is not money. It is supply of people willing to be seen taking it.

The second mode is the more useful one, because the trigger was the creator's own candour. Hank Green disclosed that he had used AI as a research tool for one of his shows and drew an immediate negative reaction from his fans ^[7]. His own account was that he had been "relying too heavily on AI as a research aid," useful for surfacing papers quickly but corrosive to his ability to find his own way into a topic ^[8]. He then published a standard: no portion of a script written, edited or outlined by an LLM, a thesis that must originate with a human, and no AI-generated image or music in a video ^[9]. That is three enumerated rules ^[10], specific enough to be tested against a finished video. It is also the only durable artefact to come out of any of these incidents.

The third mode is not governable by abstinence. As TikTok and Instagram push to detect and label artificial content, creators including Gregory Littley and Lindsey Lee Lugrin found posts promoting offline work, physical Polaroid photography and hand painting, flagged as AI ^[11]. "I cringe when I see that label," Littley said. "I cringe even more when I know it's not true" ^[12]. One of the three failure modes therefore requires no AI use at all ^[13], which means a no-AI pledge is a claim you have to be able to evidence: raw files, timestamps, process footage, and a known appeals route.

The background mood is not improving. Business Insider attributes the ramp in hostility to unpopular data centers, job displacement fears and AI-driven price rises on hardware such as game consoles ^[14], with low-quality AI video and misinformation flooding feeds ^[15].

Note the sequence in both named cases: Green wrote the policy after the reaction, Tan apologized after the post ^[4][9]. The cheaper order is to publish the standard first, then price the deal against it, with disclosure language, a right to refuse creative that reads as machine-made, and an agreed response if the comments turn.

What to watch: whether the platforms doing the labelling build a visible appeals process ^[11]; whether AI companies start paying a hesitancy premium as talent supply tightens ^[5]; and whether management firms move from case-by-case judgement to published, enforceable use policies of the kind Green wrote ^[9].