Subscribe to our FREE Newsletter, or Telegram and WhatsApp channels for the latest stories and updates.
If your social media feed has suddenly started looking like an old family photo album from the 1980s, you’re not alone.
A new AI photo trend has taken social media by storm, with people using generative AI to transform their modern-day selfies into retro portraits complete with big hairstyles, vintage clothes, film grain and the unmistakable look of an old photograph.
It is fun, nostalgic and, admittedly, quite addictive.
But there is one part of the trend that isn’t quite so retro: the enormous computing infrastructure working behind the scenes to produce those images.
So, just how bad is that AI-generated 1980s portrait for the environment?
It takes more than just a prompt
When you type a prompt and receive an AI-generated image within seconds, it can feel as though the picture simply appears out of thin air.
It doesn’t.
The image has to be generated by powerful computer processors in data centres. Those machines require electricity, while the data centres themselves require cooling systems to deal with the heat produced by all that computing.
A 2025 study examining 17 AI image-generation models found that energy consumption varied dramatically between models, with some using up to 46 times more energy than others.
The researchers also found that increasing image resolution could significantly increase energy consumption in some models.
That means there isn’t really one universal figure for how much energy your particular 1980s portrait uses.
It depends on the AI model, hardware, resolution and how the image is generated.
How much electricity does one picture use?
This is where things get a little complicated.
The United Nations University estimates that generating a typical AI image uses enough electricity to run a 10-watt LED light bulb for about 17 minutes.
That might not sound like much.
One picture certainly isn’t going to send the world’s electricity grid into crisis.
But the problem is scale.
People participating in viral AI trends rarely generate just one image. They might try several prompts, regenerate an image because their face looks slightly strange, change their hairstyle, try another outfit and then create another version for Instagram.
Multiply that by millions of people around the world and the numbers begin to add up.
The United Nations says inference — essentially the process of people actually using AI after a model has been created — now accounts for the vast majority of AI’s energy consumption as billions of interactions take place.
And image generation is considerably more energy-intensive than many basic AI tasks.
According to the same UN analysis, a typical AI-generated image can require around 1,450 times the energy of a basic text-classification task.
Then there’s the water
Electricity isn’t the only environmental concern.
AI data centres generate huge amounts of heat, and cooling them can require water.
The exact amount varies enormously depending on the data centre, cooling technology, location and electricity source, so claims that every AI picture consumes a specific number of litres of water should be treated cautiously.
But the broader problem is real.
A 2026 study published in Water Research estimated that AI’s global water footprint could reach between 4.2 and 6.6 billion cubic metres annually by 2027.
That includes water used for cooling, electricity generation and the manufacturing of the semiconductor hardware required to run AI systems.
The United Nations University estimates that the electricity-associated water footprint of a typical AI-generated image is roughly equivalent to two tablespoons of water.
Again, that is not a reason to panic over one selfie.
The concern is what happens when billions of AI interactions take place at scale.
Your one photo isn’t the problem
This is probably the most important part of the story.
If you generated an AI 1980s portrait of yourself, there is no reason to feel guilty about it.
One image is an extremely small contribution to the overall environmental footprint of AI.
The bigger issue is the sheer volume of AI-generated content being produced.
Viral trends encourage people to generate image after image, while new AI applications are increasingly being integrated into everyday services.
At the same time, companies are building more data centres and deploying increasingly powerful computing hardware to meet that demand.
The United Nations estimates that global data centres powering AI could consume as much as 945 terawatt-hours of electricity annually by 2030.
That’s nearly three times the combined annual electricity consumption of Pakistan, Bangladesh and Nigeria.
So should you skip the trend?
Not necessarily.
There is a difference between occasionally generating a fun picture and repeatedly producing hundreds of disposable AI images.
If you want to reduce your personal footprint, the easiest thing is simply to be a little more intentional.
Generate the picture you actually want rather than endlessly regenerating variations. Avoid unnecessarily high resolutions when you don’t need them, and don’t treat AI-generated content as completely resource-free just because it exists digitally.
The bigger responsibility, however, lies with the companies building and operating these systems.
More efficient models, renewable electricity, better cooling systems, recycled water and greater transparency about energy and water consumption could all help reduce AI’s environmental impact.
The viral 1980s photo trend isn’t going to destroy the environment, but it does offer a useful glimpse into a much bigger issue.
AI might feel like something that exists entirely in the digital world, but behind every generated image are physical data centres, electricity grids, cooling systems, computer chips, water supplies and natural resources.
So go ahead and enjoy your 1980s alter ego.
Just remember that while the photo may look like it came from 1985, the technology producing it is very much a product of 2026.
Share your thoughts with us via TRP’s Facebook, X, Instagram, or Threads.
Your AI 80s Photo Looks Cool — But What Does It Cost The Environment?
Entertainment Flash Report
Comments
Post a Comment