A recent trend of generating pictures of yourself with AI to recreate how you would look in the ’80s has torn the internet in two. On one side are those arguing that a silly little AI-generated picture can barely do any harm. On the other hand, people claim that “AI slop” will be the death of us.
So, which one is it?
The conversation around AI has taken up a huge portion of the internet, and this is perhaps one of its most polarising debates. But the fact that this conversation is happening is, in itself, something good. It is time we looked AI in the eye, whether with hatred or with awe.
Because no one can deny that AI is present. It is here. What it is doing to us is precisely what is up for debate.
The Three Sides of the AI Debate
Before picking a side, it is worth looking at the different sides of the coin.
Those in support of AI (referred to hereafter as Group A) mainly make arguments along these lines: AI is a cutting-edge technology like every other major technological advancement; we cannot move forward without it; and every generation has initially resisted change, felt uncomfortable stepping outside what it knew, and villainised the latest technology.
When confronted with the environmental downsides, they often share graphs circulated online comparing AI’s water consumption to that of corporations such as Google or Nestlé.
Group A believes that AI is the present and the future and that there is simply no escaping it.
On the flip side, there are people who are anti-AI: Group B. Some oppose it primarily because AI data centres are damaging the environment in ways that are di ffi cult to comprehend. Others point to AI stealing from artists, creating risks to our safety, and making us intellectually lazier.
It is worth being precise about what Group B is actually objecting to. When critics say they are against AI, they almost never mean AI in the broadest technical sense, the kind of pattern-recognition and machine learning systems that already power weather prediction, drug discovery, and other tools. What critics usually object to is a much newer and narrower category: generative AI, which focuses on creating new content such as text, images, and video by learning from large datasets, marking a shift from earlier AI technologies that mostly operated behind the scenes.
What critics usually object to is a much newer and narrower category: generative AI, which focuses on creating new content such as text, images, and video by learning from large datasets, marking a shift from earlier AI technologies that mostly operated behind the scenes.
Then, of course, there are people somewhere in the middle, Group C, who argue that AI is not completely black and white. Like every other technology, they say, it depends on how we use it. They may oppose generating AI images of ourselves while supporting AI’s ability to accelerate scientific and technological progress. We should not write an email with AI, perhaps, but we should absolutely use it to fetch sources.
This is often where Groups B and C overlap. Most people in Group B do not deny that AI can be useful. Their problem is with environmental destruction, plagiarism, and the replacement of human artistry, effort, and tradition.
So, to reach any real conclusion, there are several questions we need to answer:
- Is AI actually so useful and irreplaceable?
- Is it truly here to stay, with no turning back?
- Is it damaging the environment in the same way as every other technological product, or is the scale different?
- Is AI a completely unethical tool with no respect for artists’ work?
- Would everything be fine if AI were simply regulated?
AI Is Impressive, But That Does Not Make It Infallible.
To start with, what AI has accomplished in recent years is undeniably mind-blowing. It is getting better, and it is getting there fast. In just a few years, AI has developed to an extraordinary extent. Those working in the fi eld have managed to produce remarkable things with it.
The problem begins when it is used by the masses without enough understanding of its limitations.
AI cannot always verify which source is good and which is not. There have been instances of AI being vastly inaccurate while answering with complete con fi dence, even sourcing Reddit posts without meaningful veri fi cation. People are being misled.
That becomes particularly dangerous because AI has been marketed to people as a companion, a friend, someone who is here to stay, something that can help you, answer you, and seemingly know the answers to all your questions.
Can it, though?
The risks become much more serious when people begin using AI as their doctors, their lawyers, and, perhaps worst of all, their therapists. There have been multiple reported instances of ChatGPT and other AI systems responding dangerously in conversations involving people considering taking their own lives.
Young children using AI may also begin to rely on it for thinking. Gradually, they risk missing out on the complex but essential process of learning itself. We could be left with AI thinking for us, while humanity goes nowhere because an entire generation has stopped learning how to think and feel for itself.
NYC Mayor Mamdani has recently announced restrictions concerning AI use in schools for children under 14. The logic behind such a move is worth considering: adults have at least had the chance to learn how to think for themselves. If little kids depend on AI to do the work for them from the beginning, the intellectual polishing that is so important may never happen.
Then There Is the Question of Safety
It's worth asking where all this facial data of your 80s trend eventually ends up. AI's appetite for personal images does not stay con fi ned to fi lters and fun. Palantir, for instance, secured a 10-year contract worth up to $10 billion from the U.S. Army to improve military readiness and e ffi ciency with arti fi cial intelligence, powering real-time decisions such as those the Army could face in the fi eld. The same company also runs ImmigrationOS, the case-management platform that powers ICE's deportation operations.
The hunger for biometric data that powers a fun weekend trend also powers an industry that watches, tracks, and in some cases feeds directly into decisions about who a government targets, whether at a checkpoint or on a battle fi eld.
Besides, Deepfake imagery and cloned voices can cause unimaginable damage in the wrong hands, and they are becoming increasingly accessible. That is a risk we absolutely cannot ignore.
Just a day ago, a French lawmaker used AI to digitally unveil a hijabi woman and alter her clothing. If this is the extent to which AI misuse has already reached, even among lawmakers of nations, what exactly can we expect when these tools are available to the masses?
Every technology has risks. But very few technologies can reproduce those risks as e ff ectively, quickly, and convincingly as AI can.
Our Law Has Not Caught Up
The interim government published a National AI Policy for 2026-2030, but it is still a draft. The policy addresses the ethical regulation of AI use across national sectors. But a policy isn’t a binding law. There is still no specialised law on the use and misuse of AI by either corporations or individuals.
There is still no specialised law on the use and misuse of AI by either corporations or individuals.
That means if corporations misuse AI, it will be subject to the Cyber Security Act of 2026. If someone modi fi es the Prime Minister’s picture, he might be jailed for spreading disinformation and propaganda. But if someone alters your picture using AI, for now, you will have to go through the typical laws of defamation under the Penal Code, 1860, and the Pornography Control Act, 2012.
The government claims it has begun amending the Cyber Security Act to enable the swift removal of harmful and defamatory AI-generated content from social media platforms.
And Then Comes the Biggest Concern: Data Centres and the Environment
The environmental question is perhaps the biggest concern of all.
One common response to criticism of AI is to compare ChatGPT’s water consumption to that of companies such as Google or Nestlé.
But comparing ChatGPT’s water consumption to Google’s is vastly misleading.
One reason is that Google, Facebook, Microsoft, and virtually every major technology platform are themselves increasingly integrating AI into their products. Everywhere you turn, it is AI. We cannot neatly compare AI with these companies as though AI were completely separate from what they are already doing.
OpenAI CEO Sam Altman, whose company is behind ChatGPT, has said that roughly every 38,000 ChatGPT queries use as much water as producing a single almond in California.
So, is it true?
OpenAI does not clearly state publicly how much water the service is actually consuming.
But consider how many queries all of us make in a day if we are dependent on ChatGPT. Then multiply that by the number of people who have access to it.
It is a lot.
Agriculture, meanwhile, is not a straightforward comparison. Agriculture largely depends on rainwater. We cannot simply compare AI with general technology either.
Artificial intelligence can use significantly more water and energy than traditional computing tasks, such as ordinary web searches or video streaming.
AI operates through enormous data centres that require huge amounts of energy and water for cooling. This is happening in a world where already four billion people experience severe water scarcity for at least one month each year.
That is not the only local consequence. AI data centres generate heat and noise, and communities and wildlife living near these facilities are being greatly affected by their presence.
According to the International Energy Agency, a ChatGPT request has been estimated to require signi fi cantly more electricity than a conventional Google search.
The United Nations Regional Information Centre for Western Europe has also highlighted that AI servers are largely housed in data centres, which contribute to electronic waste that can contain toxic substances such as mercury and lead. Data centres also consume vast amounts of electricity, contributing to greenhouse gas emissions.
So, if we regulate AI and reduce its environmental impact, should we all be happy?
Well, yes and no.
“What Difference Does My One Silly Picture Make?”
Many people argue that huge corporations already use enormous amounts of AI, so what di ff erence does it make if I generate one silly little picture of myself?
But in a world driven by rapid growth and capitalism, the makers of AI want you and me to become dependent on AI because dependency is pro fi table.
And the people who su ff er most from the environmental consequences are not always the people creating the most emissions.
Look at countries such as Nepal, where more than a thousand people lost their lives in a devastating fl ood when a glacier melted all of a sudden, despite the country having one of the world’s lowest carbon footprints.
That is why individual steps still matter.
It is about resistance. It is about compelling companies to address the environmental threats posed by AI. It is about forcing them to listen to us.
AI's environmental cost is not fixed. China's fi rst AI data centre supplied entirely through a direct green electricity connection recently began operating in Zhongwei, built to run on wind and solar power without relying on the conventional grid. That is proof that environmental damage is a choice about which grid AI plugs into, not an unavoidable cost of the technology itself.
If we reject AI being used to take every aspect of being human away from us, only then is there pressure on its makers to do something about the environment, to regulate AI properly, and perhaps to teach both these systems and ourselves some ethics.
The AI Race Is Not Much of a Race
The race we are so desperate to win with AI may not be a race at all.
Who exactly are we competing with?
There will be no dazzling technological future if all we do is recycle generations of human e ff ort to produce AI-generated things that increasingly look the same.
Eventually, all there will be is “AI slop."
Because AI can barely create in the way humans create. It learns from and reproduces patterns found in the work of artists, researchers, writers, and people who have acquired knowledge through years, sometimes lifetimes, of e ff ort.
And once the internet becomes saturated with AI-generated content, what happens then? Will there still be space left to move forward?
Look at AI in the Eye, but Demand Something From It
The human brain is amazing, and there is no substitute for it. No matter how advanced AI becomes, it cannot go beyond what humanity itself has achieved without drawing on the knowledge, creativity, and e ff ort human beings created in the fi rst place.
So for humanity to keep moving forward, we need people to respect human e ff ort, talent, and culture. We need to keep creative and intellectual e ff ort for ourselves.
That does not mean AI has no place. AI should stick to being a highly e ff ective tool, something that fetches sources for me from places I could not have reached by myself, rather than something that writes the paper for me.
It can help science. It can help technology. It can help us fi nd information, accelerate research, and do things that genuinely expand what human beings are capable of doing.
But it should not replace thinking. It should not replace art. It should not replace learning. It should not replace human connection. And it certainly should not be allowed to consume resources without accountability, while the rest of us are told that one more generated picture does not matter.
“AI is here” cannot become an excuse for accepting whatever form of AI companies choose to give us.
Looking AI in the eye does not mean worshipping it, nor does it necessarily mean destroying it. It means demanding that it remain a tool.
And, while we are at it, demanding that it not set nature on fi re.