The '80s AI photo craze — clever prompt engineering or just a p***ing filter fad?

[ethanmiller]

UK social feeds are currently flooded with AI-generated 1980s portraits, built from detailed ChatGPT prompts covering everything from office headshots to arcade snapshots. As someone who follows AI image trends closely, I'm curious how this community rates the results compared to other viral photo trends we've seen.

• Out of the various setups on offer — wedding portraits, terraced house living rooms, nightclub scenes — which produced the most believable period detail for you (lighting, clothing, film grain)?
• Getting the AI to keep someone's actual likeness intact seems to be the trickiest part, since tools tend to "beautify" faces by default. What's worked best for you to stop that from happening?
• Where do you land on trends like this: meaningful creative reinvention of the past, or disposable novelty content that'll be forgotten within a season?

Share your results in the comments — I'm especially interested in which small styling touch (hairstyle, backdrop, film texture) made the biggest difference between "obviously AI" and "genuinely could p*** as an old photo." Full prompt list here if you want to try it yourself:

[SwatiSood]

I think it’s a bit of both. The results can be surprisingly convincing when the prompts include specific details like clothing, lighting and film texture. However, keeping the person’s natural likeness intact is still a major challenge. For me, it’s most interesting when used creatively rather than just as a novelty filter.

[Russell]

These 80s portraits are such a blast! I've found the terraced house living room scenes most believable, especially with that distinct yellow-tinted lighting. To keep likenesses real, I add "raw skin texture" and "unfiltered" to my prompts. It's mostly disposable novelty, but the nostalgic aesthetic is undeniably charming.