The viral 80s AI trend comes with a price

Have you already generated an image of how you’d have been in the 1980s? Did it help garner likes and comments? Generative Artificial Intelligence (AI) does just that; it makes us feel good like a high sodium content in your favourite snack. But it comes with its own risks.

The viral 80s AI trend is now a social media craze where users, for the fear of missing out, use generative AI chatbots to reimagine their modern selfies as retro portraits. While cyber security researchers scramble to warn the public about digital safety, Indian politicians are also actively normalising the 80s trend. The long list includes Union Ministers Piyush Goyal, Kiren Rijiju, Shivraj Singh Chouhan and Assam Chief Minister Himanta Biswa Sarma. Congress leader Renuka Chowdhury shared a real 1980s magazine cover of herself. Sportstars too joined in — from former cricketer Irfan Pathan to chess Grandmaster Praggnanandhaa.

Behind the retro craze lies privacy risks and environmental concerns.

Every AI generated image leaves a digital footprint. When users post high-resolution images of themselves on a third party app, this might be risky for them on the long run where users are giving away their face data. This might in turn be used in training facial recognition models without consent.

Uploading selfies grants these platforms access to biometric data and facial geometry. In 2024, there was a massive surge in cases of Deepfake scams that impersonate not only consumers but also public figures across spheres.

A bigger privacy question lurks beneath the surface: what happens to the data users provide to AI platforms?

ChatGPT, the most widely used generative AI tool, offers a revealing case study. OpenAI’s privacy policy states that data sharing is enabled by default for users in Free, Plus or Pro plans. This data would help train its models. The company uses a mix of information sources for training, including publicly available internet content, information accessed through partnerships, and information provided by users, contractors, and researchers.

Users can however, opt out of this by navigating to Settings > Data Controls > “Improve the model for everyone” and toggle the setting off. This setting is cleverly set in such a way that many users wouldn’t know what the term ‘improve the model’ means.

When India’s digital engagement is soaring, users need to be made aware of the bigger picture and look at the digital risks they are taking. A global study once concluded that Indian consumers are more prone to sharing personal information with AI agents. AI-based biometric scams are also possible and there is a lack of legal framework to govern the usage of facial recognition systems in India. It is to be noted that the Digital Personal Data Protection Act, 2023, was not drafted with generative AI in mind.

The bigger warning on warming

Millions of users generating images for fun, may not be aware of the fact that one AI image takes up as much energy as charging a smartphone. A report by TRG Datacenters in 2025 said creating an AI image uses 3 watt-hours of electricity and releases 1 gram of CO2.

It takes up a heavy cost on data centres. Compared to other types of pollution that contribute to global warming, AI image generation has a very minimal footprint. The image generation may look like nothing. The servers are all elsewhere, but AI is also a physical infrastructure, and it can double data centre power and water consumption by 2030, UN researchers had said. India’s data centre capacity is also projected to reach nearly 6.5 GW by 2030.

Data centres do not just consume power and water. They also generate heat. The collective environmental cost is a lot. Mumbai’s land surface temperature has already risen from 40.9°C in 2003 to 47.3°C in 2023, an increase driven largely by urbanisation and heat-trapping infrastructure. Chennai is also steadily emerging as a data centre hub and we know how the city has heated up considerably in the last few years. Despite climate hurdles and its recurring water woes, Chennai is a primary data center destination in India. Data centers consume millions of liters of water daily and a 100 MW data centre consumes about 2 million litres of water daily. That is equivalent to the daily use of roughly 6,500 households.

What users need to be made aware of, is that generating a single AI image requires more energy than a simple Google search. Before you upload that next selfie for a retro glow-up, remember: the trend fades in a week, but the carbon and privacy costs don’t.