When arguments get cheap

Aristotle described three ways to persuade someone. Appeal to reason, with evidence and argument. Appeal to your own credibility, so the audience trusts you enough to listen. Appeal to emotion, so they feel why it matters.

Logos, ethos, pathos. Twenty-four centuries later they still cover most of what happens when anyone tries to convince anyone of anything.

A fourth term often used alongside them is kairos: the force of timing, saying the right thing when an audience is ready to hear it.

What happens when a machine can produce polished arguments on demand, at almost no cost, on either side of any question?

What actually got cheap

Writing a well-structured argument used to cost something. Find the evidence, understand it well enough to order it, then spend real hours making the sentences carry it. That cost was quietly doing work for the reader. A polished, well-cited case was evidence of effort, and effort was weak evidence of care.

That link is now broken. A competent model can produce a structured case in seconds for almost any position, whether or not anyone has examined the evidence.

This does not necessarily make each argument worse. It makes polish a useless signal of care. The existence of a persuasive case no longer tells you whether anyone thought carefully before publishing it.

Cheap arguments meet engagement systems

Many platforms distribute content by optimizing for engagement. They did that before generative AI, when supply was still limited by human production. Social media did not invent kairos, but it industrialized it: platforms can match persuasive material to the moment when a particular audience is most likely to respond. Generative AI gives that system more material to test and amplify. Truth is not the target. The incentives have not changed, but the volume has.

One consequence is that logos gets cheap while ethos becomes scarce. If a polished argument no longer tells you whether anyone thought carefully about it, the identity, record and accountability of the speaker start to matter more.

That combination requires no villain and no breakthrough. It is an existing incentive structure with a new supply of persuasive material.

The risk is not simply a public drowning in lies. It is a public that cannot evaluate every argument on its merits and reaches for a different filter.

What history suggests happens next

History offers a useful comparison because this is not the first time a communication technology outran the institutions that checked it.

Each medium changes which appeal can scale. Print made argument and polemic cheap to reproduce, helping religious pamphlet campaigns reach larger publics. Mass newspapers rewarded pathos through yellow journalism. Radio and television concentrated ethos in a few recognizable voices and images.

The checks came later: scientific societies and peer review for published knowledge; journalism codes of ethics and regulators for mass media. In each case, expanded persuasive reach preceded institutions designed to discipline it.

What the rebuild looks like

The first verification systems are already appearing.

Content provenance standards, mainly C2PA, can attach a cryptographically signed record describing where a file came from and what was done to it.

The EU AI Act takes a regulatory approach by requiring machine-readable marking or disclosure for several kinds of AI-generated content.

The technology arrived first; verification infrastructure is following.

This infrastructure verifies origin, not truth. A provenance record can make “where did this file come from?” checkable, but not “is this claim true?” A screenshot can strip the credential, and a missing credential proves nothing. Marking synthetic content tells you how something was made, not whether it is right.

These systems still do not establish that a human stands behind an account, make a credible track record portable between platforms, or impose a consistent cost for being wrong at scale.

This is where ethos has to step in, not as blind trust in familiar people, but as accountable credibility. The next layer may make that credibility portable: proof that a person or institution stands behind a claim, a track record that follows them across platforms, and consequences for being repeatedly wrong at scale. None of this would determine whether a claim is true. It would make publishing without care more costly.

The uncomfortable part

Earlier versions of this problem were addressed by institutions, not just by more vigilant readers. Editorial standards, libel law, professional norms and regulators came from institutions, not individual vigilance.

Every new medium scaled persuasion before society learned how to govern it. AI scales logos inside systems that have already industrialized kairos. Whether ethos becomes accountable credibility or merely trust in the familiar may define what comes next.