My feed currently looks like everyone I know has travelled back to the 80s.

Big hair. Soft focus. Film grain. Power dressing. The occasional hairstyle that probably should have stayed in the decade it came from. And I love it.

There is something wonderfully simple about these AI trends. You upload a photograph, give the tool a prompt and a few moments later you’re looking at a version of yourself that never existed.

Being in the communications space, I’m naturally fascinated by anything that gets millions of people participating in the same creative idea.

But somewhere between seeing yet another 80s transformation and wondering what mine might look like, I had another thought.

We’ve become remarkably comfortable giving AI our faces. And I’m not sure we always think about it that way.

When did our face become another prompt?

Think about the transaction behind the trend. We upload a clear photograph of ourselves, sometimes several, and ask an AI platform to understand our face well enough to recreate us in another time, place or style. But that isn’t how the experience feels.

It feels like: Show me what I would have looked like in the 80s.

And that difference is interesting to me.

If a website asked us to “upload personal visual data,” many of us would probably hesitate.

Ask us to upload a selfie so we can see ourselves with an 80s blowout, and suddenly the decision feels considerably less serious. That isn’t necessarily a bad thing.

It is, however, a reminder of how quickly our relationship with technology is changing.

So, are we actually training AI with our faces?

This was the question I wanted to understand better.

And the answer is not a simple yes or no.

Different AI platforms have different policies, settings and approaches to the content we give them.

For example, OpenAI allows users of its consumer services to turn off the setting that lets future conversations contribute to model improvement. It also offers Temporary Chat, where conversations aren’t used to train its models.

Google similarly explains that, depending on Gemini settings, uploaded images can be used to improve its services, including with human reviewers. Users can change the relevant activity settings.

Adobe takes a different approach with Firefly, stating that it does not train Firefly’s generative AI models on customer content.

So saying “every AI image you create trains AI on your face” would be misleading.

But saying “it’s worth knowing what happens when you upload your face” feels entirely reasonable.

And perhaps that distinction matters.

I don’t think the answer is to stop playing

This is where I don’t want the conversation to become fearful.

I use AI. I experiment with it. I’m fascinated by what it is allowing people in creative industries to do. And I certainly don’t think the answer to every new technology is to stand at a distance until we understand every possible implication. Curiosity matters. Play matters too.

A ridiculous photograph of yourself transported back forty years might even be one of the easiest ways to start understanding what these tools can do. But curiosity and awareness can exist together.

We can enjoy the result and still spend thirty seconds understanding the tool we’re giving our photograph to.

Perhaps we need a new kind of digital instinct

We have developed certain habits almost automatically over the years. We know not to share our OTP. We think twice before entering card details on an unfamiliar website. Most of us have learned to be suspicious of the email telling us we have unexpectedly inherited $14 million.

AI is creating a new category of decisions for which our instincts are still developing. Our face. Our voice. Our writing. Our photographs. The things we create. And increasingly, the photographs and information of other people in our lives.

I don’t think that requires paranoia. Maybe it just requires a few more questions.

What am I uploading? Who am I giving it to? What does that particular platform say it can do with it? What controls do I have? And would I make the same decision if the photograph belonged to somebody else? I find that last question particularly interesting. Because participating ourselves is one decision.

Uploading photographs of our children, friends, colleagues or clients introduces someone else’s face into that decision too.

There is a communications lesson in this as well

The whole thing has made me think about language.

“Upload personal visual data for processing by an artificial intelligence system.” Not particularly inviting.

“See yourself in 1986.” We all are listening.

The action may involve the same photograph. But the way we experience the decision is completely different. Good technology increasingly removes the feeling that we are interacting with technology at all. We aren’t thinking about models, data processing or machine learning. We’re thinking about whether AI has given us the correct amount of 80s volume in our hair. That ease is part of what makes these tools extraordinary.

Perhaps it is also why our awareness has to become a little more intentional.

I hope we keep having fun with it

I don’t want a future where every interaction with technology begins with suspicion. I want the opposite. I want to see what people create. I want communications teams experimenting with ideas that would have been impossible a few years ago. I want brands playing with culture.

And yes, I want to see which decade the internet decides to send all of us to next.

But perhaps becoming comfortable with AI means becoming comfortable holding two thoughts at the same time:

This is fun. And I should understand what I’m giving it.

We don’t need to fear every new tool. We just need our awareness to evolve alongside them.