The ChatGPT Caricature Trend Might Be Fun But It Isn’t Harmless
There’s a new trend going around that looks harmless on its face: AI-generated portraits that don’t just show what you look like, but “capture who you are.” People are using ChatGPT caricatures of themselves that seems to have taken off like wildfire. The caricatures are complete with objects and symbols meant to represent their identity: their interests, their career, their family, their personality, their lifestyle.
It’s easy to see the appeal. It’s creative. It’s flattering. It feels like self-expression. It’s shareable. It’s a quick dopamine hit that on its face might seem like harmless fun - a diversion during a time of extreme stress, anxiety and trauma for so many in the US and the world right now.
But I want to say something plainly: this trend is not harmless.
Not because any single image is a catastrophe. Not because anyone who participates is a bad person. But because this is exactly how collective harm happens: one “fun little thing” at a time, until it becomes normal—until it becomes infrastructure.
And we are living through a moment where Big Tech and AI companies are grabbing more and more control over our data and daily lives. They do it through constant data collection. Through algorithms that shape what we see, what we believe, what we buy. Through founders and funders who increasingly seem comfortable making decisions for all of us—without consent, without accountability, and without serious effort to avert harm to the planet, individual rights, or humanity.
This is why I don’t think the real question is whether these AI caricatures are cool or cringe. The real question is whether we still have the awareness and self-discipline to resist trends that actually harm us and strengthen Big Tech’s hold on all of us, including our institutions, government and rule of law.
Because restraint is power. Self-control is power. And in this era, it might be the most important kind of power ordinary people still have.
Privacy: The Portrait Isn’t Just a Picture. It’s a Profile
To generate one of these portraits, people often upload a photo of their face. Then they provide personal details to help the ChatGPT “get it right”: where they live, what they do, what they love, who their kids and spouse are, what they believe, what makes them unique. Sometimes people paste in bios, resumes, or long descriptions of their life.
That’s not just a prompt. That’s allowing ChatGPT to create a profile of each individual providing the prompt. I suspect most of the people using ChatGPT for this activity are using the free version of that platform, and if so, that means that you are feeding all of this personal and personally identifying data about yourself, your family, your kids and your life, into a database that will use that information to train their AI dataset.
And in a world where data is currency—where companies have built empires by collecting, storing, and monetizing the most intimate details of our lives—it is naive to treat this as a harmless art project.
We have been trained to normalize giving away pieces of ourselves. We’ve been trained to treat privacy like it’s outdated, like it’s paranoid, like it’s not worth protecting.
But privacy isn’t about having something to hide. Privacy is about having something to keep. It’s about autonomy. It’s about safety. It’s about the right to exist in the world without being constantly tracked, categorized, and exploited.
When you tie your face and identity to these systems, you are feeding a machine that is growing more powerful and less accountable by the day.
Why ChatGPT Is Not the Place for Your Personal Life
That’s not just a prompt. It’s allowing ChatGPT to construct a profile.
When you describe yourself in detail—your face, your interests, your work, your location, your family, your children—you’re not just asking for an image. You’re providing a rich, structured snapshot of a human life. And when you do that inside ChatGPT, you are giving that information to a private company whose entire business model depends on collecting, processing, and leveraging data at enormous scale.
Most people participating in this trend are using the free version of ChatGPT. And while OpenAI has made public statements about data use and user controls, the basic reality remains this: free products are not free. You are paying with information, attention, and participation. You are contributing to the refinement and expansion of a system designed to become more powerful, more integrated, and more indispensable.
This is not about assuming bad intent. It’s about understanding incentives.
OpenAI is not a neutral public utility. It is a private company operating in one of the most competitive and politically significant industries on the planet. Its founder, Sam Altman, is not just a technologist; he is a powerful figure deeply embedded in Silicon Valley, global capital, and increasingly, government. OpenAI has active relationships with major corporations, cloud providers, and public institutions. Its leadership regularly engages with U.S. policymakers, regulators, and national security stakeholders as AI becomes a central strategic concern for governments.
That alone should give us pause.
When technology companies move this close to the state—when they position themselves as essential infrastructure for economic competitiveness, military readiness, and governance—the data they hold becomes more sensitive, not less. The stakes change. Information that feels mundane today can become consequential tomorrow, depending on who has access to it, how policies evolve, and how power shifts.
You do not need to believe in some dystopian conspiracy to recognize this risk. History is full of examples where data collected for one purpose was later repurposed in ways users never anticipated, especially once political pressure, national security framing, or corporate survival entered the picture.
The problem is not just what OpenAI says it does today. The problem is that once data exists, it can be:
used differently under new leadership,
reinterpreted under new laws,
shared under new partnerships,
or compelled under new government authorities.
And you, as an individual user, have almost no leverage once that information is already inside the system.
This is why it’s reckless to treat ChatGPT like a diary, a confessional, or a personality mirror. It is not your friend. It is not a therapist. It is not a private sketchbook. It is a platform being built at breakneck speed, under immense financial and political pressure, by people whose incentives are not aligned with your long-term privacy or safety.
Even if you trust OpenAI today, you are still trusting:
future executives you’ve never heard of,
future investors you didn’t choose,
future governments you didn’t vote for,
future policies written in moments of crisis.
And that is far too much trust to place in any company, no matter how polished its interface or reassuring its tone.
The safest assumption—the only responsible assumption—is that anything you put into a system like this should be treated as potentially durable, potentially accessible, and potentially meaningful later in ways you cannot predict now.
Which brings us back to self-control.
Just because a tool invites intimacy does not mean it deserves it. Just because a trend rewards oversharing does not mean we should comply. And just because a company asks for our data politely does not mean we owe it to them.
Restraint here is not fear. It’s wisdom. And its the most powerful form of resistance.
Doxing and Personal Safety: The Internet Connects Dots Better Than We Think
People still underestimate how easy it is to identify someone online.
Even if an AI portrait isn’t a perfect replica of your face, it’s often close enough. And when it’s paired with personal details—career, city, interests, lifestyle—it becomes a breadcrumb trail. A fingerprint.
Once you post it publicly, you’re not just sharing “art.” You’re making yourself easier to find, easier to map, easier to target.
And for women, immigrants, activists, lawyers, journalists, and anyone who has ever been harassed or threatened, this isn’t theoretical. It’s personal safety. It’s the difference between being left alone and being pursued. It’s the difference between living freely and living under threat.
The fact that some people can ignore these risks doesn’t mean the risks aren’t real. It just means they’ve been lucky.
Surveillance: Trends Like This Train Us to Cooperate
Surveillance isn’t only something that happens when a government or corporation is watching you. Surveillance is also cultural. It’s the normalization of constant exposure. It’s the erosion of our instinct to protect ourselves.
AI portrait trends don’t just create images. They train people to cooperate.
They make it feel normal—fun, even—to upload your face and personal details into systems you don’t control, built by companies you don’t know, funded by people who don’t answer to you.
What scares me is not just the portraits themselves. It’s what they represent: a society losing the ability to say no.
Because when we normalize constant disclosure, we create a world where privacy becomes socially unacceptable. Where boundaries feel inconvenient. Where self-protection gets mocked as paranoia.
That shift benefits the same people who profit from extraction.
Environmental Impact: “Just One Image” Is How Collective Harm Begins
The other thing people really don’t want to talk about is the environmental impact of their AI use.
Every ChatGPT caricature generated requires electricity. That electricity requires massive physical infrastructure. And that infrastructure consumes extraordinary amounts of energy and fresh water, while producing emissions and pollution that don’t disappear just because we don’t see them.
Is the environmental impact of one image huge? No. But that’s the point. That’s the lie we tell ourselves right before we become part of collective harm. It’s the same logic people use when they don’t vote because they think their single vote won’t matter. One action feels insignificant, until millions of people make the same choice.
Millions of people generating five, ten, twenty versions of the same image. Regenerating until it’s “perfect.” Trying different styles. Increasing resolution. Making sets of them. Doing it again tomorrow. And the next day. It all adds up quickly, and it adds up exponentially.
What makes this especially dangerous is that the environmental cost is invisible by design. AI companies do not want users to feel friction. There’s no carbon label. No water-use disclosure. No warning that says, “This computation has a real-world cost.” There’s no moment where the tool asks you to pause and consider what you’re doing.
It’s built to encourage more. More prompts. More engagement. More consumption.
That isn’t an accident. It’s the business model.
But the harm isn’t abstract, and it isn’t evenly distributed. The environmental cost of AI doesn’t fall on “the planet” in some vague, distant sense. It falls on real communities, right now, and often communities with the least power to resist it.
AI data centers require enormous amounts of water to cool their servers. In many places, that water is drawn from the same local supplies people rely on to drink, farm, and live. Communities near these facilities have already begun to experience water stress, rising costs, and shortages, while watching massive corporations consume staggering quantities of a resource that was once shared.
And it’s not just water use. The construction and operation of these data centers bring pollution into the surrounding environment—into the air, the soil, and the water. The environmental burden lands disproportionately on poorer communities, rural communities, and communities of color—places with less political leverage, fewer legal resources, and less ability to stop powerful companies from deciding that their land is expendable.
These communities don’t get the benefits of AI hype. They don’t get the glossy product launches or the stock gains. What they get is the heat, the noise, the drained aquifers, the contamination, and the long-term health risks that come with industrial-scale infrastructure planted in their backyard.
And this is the part we can’t look away from: our insatiable demand for more and more AI is what justifies the expansion of more and more data centers. Every new “fun” use case, every viral trend, every normalized act of casual consumption strengthens the argument that this infrastructure must keep growing—that more land, more water, more energy must be sacrificed.
So when we talk about environmental impact, we’re not just talking about climate change in the abstract. We’re talking about harm to other people. To neighbors we may never meet. To communities that didn’t consent to bearing the cost of our convenience and entertainment.
If we actually care about the environment, about justice, about other human beings—and not just in theory, then we can’t keep pretending this doesn’t matter.
Because “just one image” is how collective harm begins.
The Bigger Issue: The Future Is Being Built by What We Reward
Here’s what people don’t want to confront: the power we have is in what we enable.
We can argue about regulation, and we should. We can argue about ethics, and we should. But regulation is slow, and ethics statements are cheap.
Incentives are what shape technology.
And incentives come from us.
When we participate in trends like this without restraint, we teach AI companies that we will trade privacy for convenience. That we will trade environmental stability for entertainment. That we will trade long-term human rights for short-term novelty.
We teach them that there is no boundary they can’t push, because we will keep showing up and feeding the machine.
That’s how overreach becomes inevitable: not because it’s forced on us, but because it’s rewarded.
What We Can Do: Self-Control Is Collective Power
Most of us don’t control venture capital. We don’t sit on boards. We don’t own the infrastructure. We don’t get invited to the closed-door meetings.
But we do have vast power.
We can refuse to normalize harm. We can refuse to feed the beast of AI exploitation.
Self-discipline is not just a personal virtue. It’s collective power. It’s the ability to resist being manipulated by novelty. It’s the ability to stop being entertained into compliance.
And it’s solidarity, too, because opting out doesn’t just protect you. It reduces the incentive for companies to keep scaling harm, and it slows the cultural normalization of extraction.
So yes, it’s “just a trend.” But that’s the trap. The dytopian future we fear doesn’t arrive all at once. It arrives one prompt and one trend at a time.
Sometimes the most radical thing you can do is refuse.
Refuse to give them your face. Refuse to give them your data. Refuse to help normalize a future where the environment is collateral damage, privacy is obsolete, and the self is just another product.
If we want a different AI future, one that actually takes harm seriously, one that protects people, one that respects the planet, then we can’t just criticize the companies while continuing to feed them with our attention and data.
We have to start acting like responsible AI users now. We have to demand responsible AI now.
Because the future is being built by what we reward.
And we still get to decide what we reward. Let’s use our power for good.