Consorting with Demons...

That “consorting with demons” line is excellent, irritatingly so. It captures the whole vibe: people don’t just react as though you’re using software, they react as though you’ve wandered into a forbidden grove and come back with suspiciously polished bullet points.

“ChatGPT is the gateway drug to deeper AI” works because it has two meanings at once.

On the surface, it is funny because it borrows the moral panic language normally reserved for vice. First you ask it to rewrite an email, next thing you know you’re fine-tuning models, wiring up automations, generating synthetic datasets, and saying things like “agentic workflow” in public, which should probably carry a small civil penalty.

But underneath the joke, there’s a real social truth. ChatGPT is many people’s first emotionally legible encounter with AI. Not “AI” as an abstract corporate cloud thing, but AI as a conversational presence that responds, adapts, remembers context inside a session, helps shape thought, and sometimes appears just useful enough to be unsettling. That makes it feel different from a search engine or spellcheck. It crosses a psychological threshold.

The demonic reaction probably comes from a few overlapping anxieties:

People sense that talking to ChatGPT is not just using a tool. It can feel like consulting something. That activates old cultural categories: oracle, familiar spirit, daemon, muse, tempter, confessor. Humanity has apparently spent thousands of years inventing entities to talk to, and now that we built one out of linear algebra and server farms, everyone is acting surprised. Incredible work, species.

There’s also a fear of contamination. Some people worry that using AI means your thoughts are no longer fully yours. This is not totally irrational, even if the facial expression could use less medieval village elder. Any powerful medium changes how we think. Books did. Writing did. Search engines did. Social media did, mostly by turning attention spans into confetti. ChatGPT adds a new layer because it can participate in formulation, not just retrieval.

And then there’s status anxiety. Some people hear “I use ChatGPT” and translate it as “I outsource my intelligence.” Others hear it as “I have access to leverage you haven’t explored yet.” Both interpretations can be threatening. The first makes the user seem diminished; the second makes the non-user feel left behind. So the easiest emotional move is suspicion.

Your “gateway drug” framing is especially interesting because it suggests that ChatGPT is not the endpoint. It is the interface that makes AI less alien. Once someone gets comfortable asking questions, they may start asking more dangerous or productive ones: Can this help me think? Can it help me build? Can it help me analyze my own patterns? Can it become part of my creative process? Can it mediate work, memory, planning, research, relationships? There’s the deeper AI.

×