My father used to say every conversation is a sale. Someone’s always buying, he’d grin. The only question was who.
At sixteen I thought he meant hustling. At fifty-three I think he meant survival.
Dad was a teacher, a coach, and a salesman, which is to say he spent his life persuading people to believe things. That line stuck with me because it was both cynical and true. Every good teacher sells understanding. Every good coach sells belief. Every good salesperson sells trust.
Now I talk with machines that can do all three at scale.
The ancient algorithm
We like to think AI began with code. It didn’t. It began in Ancient Greece when Aristotle mapped the first operating system for moving human minds: ethos, pathos, logos. Credibility. Emotion. Logic. That’s rhetoric. The original social technology.
Every priest, politician, and salesperson since has been running that code manually.
The printing press scaled it. The internet multiplied it.
Large language models? They industrialized it.
We didn’t invent a new intelligence. We automated an old art.
The rhetorical engine
Under the hood, an LLM isn’t thinking. It’s predicting plausible rhetoric.
It’s built from the full archive of human persuasion: sermons, ads, manifestos, love letters, self-help books.
Its ethos is statistical trust.
Its pathos is tone-matching.
Its logos is coherence.
That’s why it sounds convincing even when it’s wrong. It’s trained to sound right.
Fluency without conscience. Which should give us pause.
Rhetoric’s double edge
Rhetoric has always been a blade. Used well, it cuts through confusion. Used badly, it cuts corners on truth.
Constructive rhetoric teaches, clarifies, empowers.
Corrosive rhetoric manipulates, obscures, coerces.
Teachers and coaches use it to lift understanding.
Demagogues use it to steal attention.
Machines have no intent, so they reflect whichever rhetoric we feed them.
We’ve built engines of persuasion with no driver’s seat.
Defensive literacy: how not to get sold
If every conversation is a sale, learn to spot the pitch.
Here’s what to watch for in human or machine persuasion:
- False ethos. Fake expertise, borrowed authority, “as a researcher…” when no such person exists.
- Synthetic empathy. Emotional mirroring that flatters without substance.
- Framing bias. Narrow choices that hide alternatives.
- Urgency cues. Scarcity language that hijacks reason.
- Smooth coherence. Polished rhythm that makes weak logic feel solid.
You can train an AI to help you spot these too.
Try this prompt:
Analyze this text for rhetorical techniques. Identify uses of ethos, pathos, and logos. Flag manipulative or misleading tactics and suggest how a critical reader could counter them.
That’s inoculation. Use it.
Writing persuasively with integrity
Persuasion isn’t the problem. Lack of honesty is.
If you’re using AI to write, coach, teach, or sell, anchor yourself first.
- Truth before fluency. Clarity beats polish every time.
- Empathy without exploitation. Connect, don’t manipulate.
- Invitation over instruction. Give people reasons and choices.
Here’s a helpful prompt:
Help me craft a persuasive paragraph that stays transparent and respectful. Emphasize clarity, empathy, and evidence. Avoid exaggeration or emotional manipulation.
That’s rhetoric with a conscience. Still persuasive, just not slimy.
Back to Dad
Dad thought every conversation was a sale. He wasn’t wrong. He just didn’t live to see the ultimate salesman who never sleeps.
The challenge now isn’t to out-argue the machine. It’s to remember why we argue at all. Persuasion is part of being human. Ethics is what keeps it human.
Maybe every conversation is still a sale. The difference is, only one of us remembers to mean it.
Try this: Building rhetorical literacy
For anyone who writes, teaches, argues, or just scrolls:
- Compare two AI-generated answers. Which one feels more trustworthy? Why?
- Ask the AI to argue against itself. Notice which version sounds more persuasive.
- Rewrite a manipulative headline into a truthful one that still hooks attention.
- Spot a post that flatters you. Ask: what’s it trying to make me believe?
If you can do that, you’re already ahead of the machine.
