Published May 22, 2023 · By CampaignsLive · Insights
In late March 2023, an image of Pope Francis in a long white Balenciaga puffer coat, hands tucked into the pockets, walking through a city street with the calm composure of an off-duty celebrity, went viral on Reddit, Twitter, and Instagram. The image was striking, plausible, and almost universally accepted as real on first viewing. It was generated using Midjourney v5 by an anonymous user in Chicago.
The image was not a brand campaign. The episode is in this archive anyway, because more than any case study from the previous year, it taught brand teams what generative imagery had become and what was about to change.
What the moment actually proved
Three things at once.
The first was that AI image generation had crossed a quality threshold the previous twelve months of brand-AI experiments had not made visible. The Pope image was not striking because it looked like AI; it was striking because it did not. The compositional plausibility, the natural light direction, the texture of the coat fabric, the proportional anatomy of the figure — all of these passed casual inspection. The image was, for nearly all viewers on nearly all platforms, simply a photograph until told otherwise.
The second was that the audience’s default trust in photographic imagery, which had held for nearly two hundred years, was no longer reliable. A decade of Photoshop manipulation had shifted that default slightly. The Pope image shifted it considerably more. Within a few weeks of the image’s circulation, news outlets, social platforms, and individual users were operating with a meaningfully higher baseline of skepticism toward unattributed photography.
The third was that the cost of producing a convincing fake image had collapsed. The Pope image was made by an individual in a few minutes using publicly available tooling. The previous threshold — a skilled retoucher, hours of work, professional reference material — had been replaced by something an order of magnitude cheaper and faster.
Why this mattered for brand creative teams
The episode did not directly affect any specific brand’s marketing program. The relevance to brand creative was second-order, in three ways.
Audience trust assumptions had to change. Brand creative had operated for decades on the assumption that the audience would, on first viewing, accept a photograph as documentary evidence of the depicted scene. After March 2023, that assumption could no longer be relied on. The audience now arrives at photographic brand imagery with a baseline level of “is this real” skepticism that did not previously exist.
The disclosure norms accelerated. The Marvel Secret Invasion controversy, the Levi’s-Lalaland.ai backlash, and the broader conversation about when AI-derived imagery should be disclosed all sharpened in the weeks following the Pope image. The cultural moment created the conditions in which “did you use AI?” became a reasonable question to ask of any new piece of striking visual content.
The competitive dynamics in image tooling shifted. The Pope image was made with Midjourney v5. Within months, the major image-generation platforms were competing aggressively on the kind of photographic plausibility the image had demonstrated. The pre-March-2023 visual register — the slightly stylized, slightly painterly default look of AI imagery — gave way to a much more photographic one across the major platforms.
What followed in the next six months
The Pope incident sat at the start of a six-month stretch in which the public conversation about generative AI shifted from “interesting novelty” to “consequential cultural force.” A partial list of what happened in the immediate aftermath:
- The “Trump arrest” AI images, circulated in late March 2023 alongside the Pope image, demonstrated that political-figure deepfakes were now within reach of any user.
- The fake “Heart on My Sleeve” Drake/Weeknd song, released in mid-April 2023, brought the same conversation into music.
- Christopher Nolan, Tom Hanks, and several other prominent figures publicly named AI synthesis as a concern in their industries.
- The Writers Guild of America strike began in May, with AI use in script generation as one of the named issues.
- The SAG-AFTRA strike followed in July, with AI clauses central to the negotiation.
- Italy temporarily banned ChatGPT over data-protection concerns; the broader EU AI Act process accelerated.
The aggregate effect was that the regulatory, contractual, and cultural environment in which brand creative would be produced for the next several years was being set up in the months following March 2023. Brand teams that did not pay attention to that period are now operating in a world they did not see being built.
What this implies for brand-AI posture
Three working implications survived the moment.
Disclosure is no longer optional. The cultural default has shifted enough that audiences expect, in many contexts, to be told when content is AI-derived. Brands that omit the disclosure are not getting away with anything; they are absorbing a slowly-accumulating reputational risk that surfaces over the course of months rather than days.
Photographic plausibility is no longer evidence. Brand creative cannot rely on an image looking real to communicate that the depicted scene actually happened. If the audience needs to believe the scene happened — a real customer, a real product test, a real moment of brand experience — the work has to be photographed, with appropriate documentation, in a way that is itself defensible.
The trust signal moved. The signal that distinguishes documentary brand work from generative brand work is no longer the visual register of the image itself. The signal is the provenance — who made it, with what tools, on what date, with what documentation. Brand teams that build production processes around documenting that signal are operating with the cultural reality the Pope image made visible. Brand teams that do not are continuing to operate as if it were still 2022.
The longer arc
The Pope-in-a-puffer-jacket image is, in retrospect, one of the cleanest single moments to point to as the start of the cultural conversation about generative AI’s implications for visual media. The image was harmless. The cultural shift was not. For brand creative teams, the practical effect was to compress a five-year transition into a one-year window. The norms, contracts, disclosure expectations, and audience defaults that had been gradually accumulating were suddenly all present at once, and brand-side production processes had to catch up.
By mid-2025, most of that catch-up had happened — at least at the leading brands and agencies. The brands that did not adapt in 2023 and 2024 spent 2025 working through the same questions late, with less leverage. For more on how disclosure norms emerged through this period, see Marvel’s Secret Invasion AI Title Sequence. For the labor side that crystallized in the same window, see The SAG-AFTRA AI Clauses.