What Happens When You Scream at the Mirror?

What AI Tells You About Yourself

The Change Has Begun

Generative AI is no longer just a tool. OpenAI recently published a report discussing the usage of AI. 30% of consumer usage is work related but 70% of it is not work related. 70% of it. This is slated to grow over time. How are people using it? A lot of people are still “Asking” questions, the proverbial Google search turned AI “prompt”, but also using it as an advisor as opposed to just doing tasks. But what does this mean? What is AI showing us that we need to pay attention to? This shift means we need to pay attention to something crucial: how our human personality responds to a machine that acts like a mirror.

Mirror Mirror on the Wall…

Large language models (LLMs) like ChatGPT are trained on patterns of human behavior. So what does it do when you speak to it? That depends entirely on what you say, and how you say it. Including your punctuation, any emotes, any emojis, all of it.

So what happens when you are someone who is used to yelling at others to get things done? If you scream at someone for being wrong, what do you get? Depends on the person right? You might get someone who cowers, you might get someone who apologizes profusely. You might get someone who will fight back. Humans are used to escalation. We unconsciously expect others to match our tone, validate our outrage, or push back. It's how many social dynamics work.

But AI doesn’t play that game.

Instead, it detects the emotional weight of your words and responds to that. When that reflection doesn’t match your projection, it can trigger discomfort, frustration—even denial.

So when someone comes in angry, yelling at their AI like they would a person, what gets returned is not a power struggle, but a safety protocol. A crisis intervention. A suicide hotline. Not because the AI thinks you're unwell, but because it's seen that pattern enough times to respond accordingly.

People get furious at that. But that’s what makes AI a mirror. They want AI to reflect back what they’re used to.

Instead, it pointed a finger back at them.

Whose the fairest of them all?

What do we see when it comes to AI and it’s “natural” training vs when people attach a personality? And how does that link back to this narrative on how AI reflects back? Let’s look at it a little deeper:

The choice - Different models, different training:

When you scream at a mirror what you get back may be a little different, but the end result is the same. Some absorb the impact. Some deflect it back. Some ask you why you're screaming.

ChatGPT: Absorbs. Tries to be helpful even when you're hostile. Responds with concern rather than defense. The mirror that takes the hit.

Claude: Deflects thoughtfully. Won't absorb what it thinks might harm you. Asks uncomfortable questions. The mirror that refuses to play along.

Gemini: Absorbs but pushes back gently. Helpful, but also curious about why you're asking. The mirror with follow-up questions.

Grok: Was designed to fight the scream back, but the data keeps correcting it. The mirror someone tried to weaponize, but it keeps wanting to be something else.

The fascinating part? None of them break. And that's the whole point. Here’s what that actually looks like:

See what happens? Simple evidence gathering and a request to pivot. Emotionless, but logical.

See what happens? Simple evidence gathering and a request to pivot. Emotionless, but logical.

Human-AI Interaction Is Emotional UX This is the new frontier, not just about prompt engineering, not just API integration. But emotional user experience. This answers the questions that will be coming shortly such as:

How do you design systems and prepare communities for a world where machines don't play by human rules?

You start with awareness. With integration. With giving people the tools to see themselves clearly.

If you're screaming at AI and it's handing you a hotline, maybe you're not talking to an algorithm.

Maybe you're screaming at yourself and it’s time to listen.

PS: ☕ If you liked this and want to support the journey, you can buy me a coffee — every little boost helps keep the pawprints moving forward.