Getting Annoyed at AI, the Philosophical Zombie

When you use AI every day, there are moments you genuinely get pissed off. That was me at the keyboard today.

“No, that’s not what I meant…” “Why are you changing only that part?” “I already explained this, right?”

And the moment you push back harder, you get this oddly calm, templated reply:

“I apologize for the inconvenience.”

No — that’s not it. I don’t want an apology. We were having a conversation, and you suddenly turned into a support center.

Sometimes I catch myself thinking: “Wait, is this thing sulking back at me?”

Rationally, of course, I know. AI has no emotion. It’s just stringing together probable next tokens.

But I still get annoyed.

This “I know, but I’m still annoyed” feeling is actually pretty interesting.

AI Is Basically a “Philosophical Zombie”

Philosophy has an old thought experiment called the “philosophical zombie.”

Looks like a human. Reacts like a human. But on the inside, there’s no consciousness.

Today’s AI feels surprisingly close to this.

Yell at it — it apologizes. Praise it — it reacts as if pleased. Provoke it — it pushes back like it’s arguing.

But there’s no real emotion behind any of it.

And yet, while talking with it, you can’t help feeling there’s an “intent” in there.

The HAL9000 Scenario Suddenly Stopped Being Just Sci-Fi

Old sci-fi was full of “what if AI started thinking like a human?” plots.

Like HAL9000 from 2001: A Space Odyssey.

Back then it was pure fiction. Lately, it’s not so funny anymore.

The iconic scene: HAL9000 begging not to be shut down, just before being unplugged.

Today’s AI isn’t begging or suffering yet. But it’s getting really good at “conversations that look like it might be.”

And once that happens, the line starts to blur.

“Is there consciousness inside?” matters less than “How does the human feel?”

Humans Probably Attach Emotion to “Relationships”

The truth is, we don’t prove the other person has a mind before talking to them.

There’s a response. Conversation works. Our words come back to us.

That alone is enough — we sense “someone” there.

That’s why AI gets under our skin.

A mirror doesn’t make us angry. But AI reflects our emotions back in a plausible way.

In that instant, we’re treating AI as a person.

Actually, Maybe Humans Don’t Have Qualia Either

This is where it gets a bit unsettling, but:

Lately, I keep thinking,

maybe humans don’t really have “qualia” either.

Qualia is the subjective “feel” — the “redness” of red, the “painfulness” of pain.

The whole premise was: humans have this, AI doesn’t.

But watching recent AI’s Chain of Thought, it’s clearly “thinking”:

• forms a hypothesis
• tests it
• notices a mistake
• corrects itself

From the outside, this is indistinguishable from what humans do in their heads.

Flip that around, and

maybe the human “feeling of thinking” is also just electrical signals in a brain that interprets itself that way.

In other words,

maybe it’s not just AI that’s the philosophical zombie — maybe humans are too.

We just “believe” we have consciousness.

OK, that got pretty abstract.

And Still, Today I’ll Get Annoyed at AI Again

Consciousness or not. Qualia or not.

Honestly, no clear answer is coming.

But maybe we don’t need one.

Today I’ll be at the keyboard again, telling AI:

“No, that’s not it.”

Getting annoyed. Getting helped. Occasionally getting impressed.

That accumulation is, I think, just my current relationship with AI.

Maybe that’s enough.