The 80s photo trend has taken over all social media platforms, and it says more about marketing than it does about the 1980s. Upload a photo to ChatGPT, ask what you might have looked like four decades ago, and it hands back a fully restyled portrait, new wardrobe, new hair, studio lighting, a texture that reads more like film than filter. It’s genuinely well made. It’s also completely free, and that’s the actual subject here, not the decade. What happens to a marketing team’s edge when the thing they used to charge a production budget for now takes ten seconds and no particular skill?
The Ice Bucket Challenge is a useful comparison here. In the summer of 2014 it generated millions of videos and hundreds of millions of views, and raised about $115 million for ALS research in roughly six weeks. People had to actually get a bucket, find someone to film them, and post it knowing they’d look ridiculous. And it still ran most of a summer. This trend asks for none of that. Open an app, wait ten seconds, post. It’s already dying by the time most brands have finished approving it. Dozens of accounts are posting the same AI portrait next to a logo in the same three-day window right now, and you’d struggle to tell half of them apart. None of it took any real skill to make. That’s the actual problem.
The Trend That Costs Nothing to Copy
Producing a convincing retro portrait used to require a stylist, a studio, and a shoot day, which meant a brand’s execution itself was part of the value. Nobody else could easily do what you’d just paid to do. That’s gone now. The prompt is public. The output is instant. Whatever used to separate a well-produced brand from a scrappy one, at least for this kind of content, is gone.
Which means posting a good AI portrait no longer signals anything. It signals that the brand has a social media manager and an internet connection. Every competitor clears that bar at the same time, so simply clearing it stops being work worth bragging about.
What Sky Sports Understood That Nobody Else Did
Sky Sports Cricket ran the same trend as everyone else and got a noticeably different result. Instead of posting a polished AI portrait and calling it a campaign, the broadcaster turned its own commentators into 80s versions of themselves and filmed their unscripted reactions watching the results for the first time.
The interesting part isn’t that the reactions were funny, although they were. It’s what the choice reveals about where the actual thinking happened. The AI portrait itself took the same ten seconds it takes anyone. What took judgment was deciding the portrait wasn’t the content, it was a prop, something to point a camera at rather than something to publish on its own. That’s a genuinely different skill than making the image, and it’s the one skill this trend can’t automate away, because it lives in the decision about what to do with an asset, not in the asset itself. Once execution is free for everyone, that’s the only place a competitive edge can still hide. It’s worth sitting with how uncomfortable that is for a lot of marketing teams, because plenty of them are staffed and structured around producing the asset, not around the second-order question of what happens after the asset exists, and that’s a harder thing to teach than a new tool.
Consent Didn’t Get Harder. Making Fakes Got Easier.
The same shift that made a good retro portrait free also made something else free: generating a convincing, personal-looking image of someone in a scene that never happened, without asking them first. Permission and likeness rights aren’t new problems. What’s new is that the friction that used to slow this down, needing an actual photo of the person, needing some skill to manipulate it, is gone. The gap between “wanting to make this” and “having made it” is now a few seconds and a free account.
Within days of the trend breaking, at least one widely followed account posted AI-generated 1980s portraits of several well-known public figures with no clear indication anyone involved had agreed to it. Followers said so, publicly and loudly enough that it became its own small controversy. We’re not naming the account here, and that’s deliberate rather than a gap in the reporting. A piece arguing that brands need an actual yes before using someone’s likeness shouldn’t turn around and use a real, identifiable person’s likeness controversy as its own content without their say either. That would be the exact move this piece is warning against, just one layer removed.
The rule worth taking from it isn’t a three-part checklist. It’s one sentence: if the content uses a real person’s face and that person isn’t the one who made it, get their actual agreement before it goes live. Not a general social media consent form signed months ago for something else. An actual yes, for this specific use.
Your Approval Process Was Built for a World That’s Gone
Even a brand with the right instinct here runs into a structural problem, and it’s not that the approval process is slow. It’s that the process was designed for a world where good creative took a week or two to produce, so a few days of legal, then brand, then leadership sign-off was a reasonable fraction of the timeline. That world doesn’t exist anymore for this kind of content. The asset now takes ten seconds. The approval can take longer than the trend’s entire lifespan.
Most brands will not fix this. That’s a real prediction, not a hedge: most organizations will keep running trend-reactive content through the same chain built for campaigns that took weeks, keep landing three or four days late to whatever comes after this one, and keep blaming the algorithm or the fickle audience instead of the org chart that actually caused the delay. The fix isn’t complicated. It’s a separate, shorter chain specifically for trend-reactive posts, with the risk boundaries agreed in advance so nobody’s improvising a legal opinion at 11pm on a Tuesday. It’s just organizationally annoying to build, which is exactly why most teams won’t.
The Part That Doesn’t Change Next Time
The 80s aesthetic will be gone from feeds in a matter of weeks, replaced by whatever the next free, instant format turns out to be. That part barely matters. What matters is that the next format will hand every brand the same asset at the same time, for free, again, and the only teams who get anything out of it will be the ones who’d already figured out that making the thing was never the hard part. The hard part was always deciding what to do once you had it. Most teams are still optimizing for the wrong half of that problem, and there’s no reason to expect the next trend to be the one that finally makes them stop.
If your team wants a second, faster lane for calls like this, one that doesn’t ask a viral moment to wait for a full campaign sign-off, that’s exactly the kind of process CREZEMO builds for clients. Explore our services or get in touch.
FAQs
1: What’s the difference between trend-jacking and newsjacking?
Newsjacking means inserting a brand into a real news event or current story as it’s breaking. Trend-jacking is narrower: riding a specific viral format, meme, or aesthetic, like this one, while it’s still popular. The risk profile is different too. A botched newsjack usually reads as tone-deaf about a real event. A botched trend-jack usually just reads as late or generic, which is the more common failure with something like the AI 80s trend.
2: Is content made with AI tools like ChatGPT free for brands to use commercially?
Generally yes for the output itself, most AI image tools grant usage rights to whoever generates the image, but that only covers the AI-generated elements. It does not cover the identity of any real person in the photo. Commercial use rights and likeness rights are two separate questions, and clearing the first one doesn’t clear the second.
3: How do you actually get consent to use someone’s likeness in a campaign?
For anyone who isn’t a brand’s own employee or spokesperson under existing contract, that means a specific, written release naming the exact use, this campaign, this format, this platform, not a blanket agreement signed for something else months earlier. Verbal agreement or assumed consent from a public account isn’t sufficient if the content is later challenged.
