• home
• about
• new
• essays
• AI
• book

Frank Visser, graduated as a psychologist of culture and religion, founded IntegralWorld in 1997. He worked as production manager for various publishing houses and as service manager for various internet companies and lives in Amsterdam. Books: Ken Wilber: Thought as Passion (SUNY, 2003), and The Corona Conspiracy: Combatting Disinformation about the Coronavirus (Kindle, 2020).

Testing the Six Strongest Claims in the AI-Johnstone Response

Frank Visser / ChatGPT

The simulated Johnstone response contains several genuinely strong arguments. It would be easy to dismiss it because it was generated by AI, but that would miss the interesting part. The response succeeds because it draws upon real historical failures of Western journalism and combines them with a coherent critique of power. The question is whether those legitimate premises actually establish the stronger conclusions the response draws from them. Here are six of its strongest claims.

1. “The New York Times, BBC and CNN have committed serious failures, so why single out independent journalists?”

This is the strongest challenge to my original essay. The AI response points to the New York Times and its role in amplifying the case for the Iraq War, and to the broader dependence of Western journalism on government and intelligence sources. It argues that when establishment media launder propaganda into consensus reality, this is treated as an unfortunate institutional failure, whereas independent journalists are suddenly subjected to intense demands for methodological rigor.

Validity: High.

There is no serious dispute that mainstream journalism has sometimes failed spectacularly. The Iraq War is an especially devastating example. It is entirely legitimate to ask why a journalist such as Johnstone should receive more scrutiny for challenging official narratives than the institutions that helped establish those narratives in the first place.

Evidence: Strong, but incomplete.

The historical evidence for mainstream-media failures is substantial. But the argument commits a subtle substitution. The question “Why don't you criticize the New York Times?” is not evidence that Johnstone is reliable. It is evidence that the New York Times has problems.

That distinction is crucial.

If I were claiming that mainstream journalism is reliable and Johnstone is uniquely unreliable, the criticism would be devastating. But I am not. My position is that both should be scrutinized.

Rhetorical sleight of hand: Whataboutism disguised as proportionality.

The AI response turns a criticism of Johnstone into an apparent defense of the New York Times. But criticizing one does not imply defending the other. The real alternative is not “mainstream media or Johnstone.” It is skepticism toward both.

The argument therefore succeeds as a criticism of insufficient media criticism but fails as a defense of Johnstone's particular epistemic practices.

2. “If your house is on fire, it isn't hypocrisy to spend more time discussing your own roof than your neighbor's.”

This is probably the most intuitively persuasive metaphor in the response. Johnstone is a Western writer addressing Western audiences. She lives inside the American-led information environment and has vastly more opportunity to influence her own society than Russian, Chinese or Iranian audiences. Therefore, why should she be expected to spend equal time criticizing their propaganda?

Validity: High.

This is a legitimate point, and I would concede it more strongly than I did in the original essay. A journalist does not have a moral obligation to distribute attention equally among all political actors. A British journalist investigating British government propaganda is not required to publish an equally long investigation into Chinese propaganda before every article can be considered legitimate.

Evidence: Essentially irrelevant to the core issue.

The geographical and political location of the journalist explains her choice of subject matter. It does not establish whether her analysis of that subject matter is correct.

Rhetorical sleight of hand: confusing attention with skepticism.

This is where the metaphor breaks down.

I am not arguing that Johnstone should spend equal amounts of time criticizing Washington and Moscow. I am arguing that when she discusses Moscow, Beijing, Tehran or Hamas, she should apply the same epistemic standards that she applies to Washington.

There is nothing hypocritical about spending 90 percent of one's time investigating one's own government. There is a potential epistemic problem if the remaining 10 percent is treated according to a different standard of evidence.

The house-on-fire metaphor therefore successfully answers a criticism I wasn't actually making.

3. “America is the empire. That's not a slogan. It's an empirical description.”

This is the most ambitious claim in the response. The AI-generated Johnstone argues that America's worldwide military, economic, financial, technological and cultural power makes the designation “empire” a straightforward empirical fact. Consequently, concentrating on American power is not ideological bias but proportionality.

Validity: Partly high, partly dependent on definitions.

The United States is unquestionably the world's dominant military and financial power, with a worldwide alliance network and extensive overseas military presence. There is ample historical scholarship using concepts such as informal empire, hegemony and imperial power to describe aspects of American foreign relations.

But “America is the empire” is not merely an empirical observation. It is also a conceptual interpretation.

Does an empire require territorial annexation? Political subordination? Economic domination? Military bases? Unequal sovereignty? Informal influence? Cultural hegemony? Depending upon the definition, the empirical claim changes.

Calling America “an empire” can therefore be defensible. Calling it an entirely neutral description is harder to defend.

Evidence: Strong for American predominance; weaker for the rhetorical conclusion.

The United States is extraordinarily powerful. That is not controversial. But power does not automatically establish imperial intent in every individual foreign-policy decision.

Rhetorical sleight of hand: turning a contested interpretation into a fact.

The phrase “That's not a slogan. It's an empirical description” is itself a rhetorical move. It closes an argument by declaring the conceptual classification to be empirically settled.

But even if we accept “American empire” as an appropriate description, another question immediately arises: does being an empire make everything said by its geopolitical opponents more credible?

Obviously not.

And that is where the argument loses its force.

4. “My side is ordinary human beings. I don't have a side.”

This is perhaps the most philosophically attractive statement in the AI response. Johnstone is represented as rejecting the idea that she is pro-Russian, pro-Chinese, pro-Iranian or pro-Hamas. Her allegiance is supposedly to ordinary human beings rather than governments.

Validity: Morally admirable; epistemically insufficient.

Of course one can oppose all governments when they commit atrocities. One can be simultaneously opposed to American imperialism, Russian aggression, Chinese repression and Hamas terrorism.

But “I have no side” does not mean “I have no perspective.”

Every writer selects subjects, sources, historical frameworks, metaphors and explanatory models. Johnstone's recurring conceptual vocabulary—empire, propaganda, militarism, NATO, censorship, intelligence agencies and corporate media—is not accidental. It represents a coherent worldview.

That does not make it wrong.

But it makes her situated.

And situatedness is not an insult. It is a basic fact of human cognition.

Rhetorical sleight of hand: confusing moral universalism with epistemic neutrality.

One can have universal moral sympathies while still possessing very strong political and epistemological commitments.

“I care about ordinary human beings” tells us something about one's moral orientation.

It tells us rather less about whether one's interpretation of a particular war is correct.

5. “I'm not replacing one narrative with another. I'm exposing how narratives are manufactured.”

This is perhaps the most interesting claim of the entire response. The simulated Johnstone says that she is not primarily interested in deciding which geopolitical narrative wins. Her task is to expose the machinery that manufactures narratives: governments, intelligence agencies, anonymous sources, think tanks, defense contractors and media organizations.

Validity: Very high as a journalistic aspiration.

This is exactly what investigative media criticism ought to do. Political narratives are constructed. Sources have incentives. Institutions have interests. Information does not simply arrive in the public sphere untouched by selection and framing.

But there is an important problem.

Narrative analysis and truth determination are not the same activity.

Discovering that an institution has an incentive to promote a particular story does not demonstrate that the story is false.

The CIA can have an incentive to exaggerate a threat and still sometimes accurately report one.

A defense contractor can benefit from a war and still occasionally provide accurate information about a military capability.

A Russian journalist can have an incentive to oppose NATO and still produce excellent investigative reporting.

A Western journalist can have access to powerful sources and still expose genuine wrongdoing.

The existence of an incentive is evidence about the context of a claim, not necessarily about its truth value.

Rhetorical sleight of hand: moving from “who benefits?” to “who is telling the truth?”

“Who benefits?” is a good question.

It is not the only question.

The other questions are: What happened? What is the evidence? What alternative explanations exist? What would falsify the claim? What evidence would change my mind?

Without those questions, media criticism can become an endless exercise in revealing motives while never establishing facts.

6. “The establishment has a flood; independent media has a garden hose.”

This is the most memorable argument in the response. The establishment narrative has governments, intelligence agencies, corporations, universities, think tanks, Hollywood and the legacy press behind it. Independent journalists have little more than newsletters. Therefore, treating the two as equivalent epistemic dangers is absurd.

Validity: High concerning power; low concerning epistemology.

The institutional asymmetry is real. There is no equivalence between a government-supported information ecosystem and an independent blogger with a Substack account.

But that is not what my argument claimed.

I did not say that Caitlin Johnstone has the same propaganda power as the Pentagon. Obviously she does not.

The relevant question is much smaller: Can a small independent media ecosystem develop its own biases and epistemic distortions?

Of course it can.

A garden hose does not need to become a flood before the water coming out of it can be contaminated.

And there is another important consideration. The danger posed by a source is not simply a function of its institutional size. A small publication can sometimes have enormous influence when its arguments are amplified through social media, podcasts, political movements or larger media organizations.

So the flood-versus-garden-hose analogy is rhetorically brilliant but logically overextended.

It successfully establishes asymmetry of power.

It does not establish asymmetry of truth.

What survives the test?

After stripping away the rhetoric, the AI-generated response leaves us with a position I actually find quite reasonable.

Western media deserve much more scrutiny than they often receive. Governments systematically manipulate information. The United States has enormous geopolitical power. Independent journalism performs an important corrective function. Political journalists should investigate incentives and institutional interests. And there is indeed something suspicious about demanding that critics of Western power constantly reassure their readers that Russia, China or Iran are also bad.

I accept all of that.

But none of these propositions establishes the conclusion that the AI-generated Johnstone response wants us to reach.

The missing step remains the same.

Skepticism toward the establishment is not evidence for the counter-narrative.

That is the entire point of my original criticism.

And perhaps the most revealing feature of Dillard's AI experiment is that the simulated Johnstone response is so good at defending itself precisely because it uses arguments that are almost impossible to disagree with at the level of general principle. Of course governments lie. Of course media organizations have biases. Of course the United States is powerful. Of course independent journalism is valuable. Of course propaganda exists.

But the hard intellectual work begins after those propositions have been accepted.

What actually happened?

Which claims are true?

Which are false?

Which sources deserve confidence in this particular case?

What evidence would change our minds?

And can we apply the same standards when the conclusion is politically inconvenient to our own preferred narrative?

That last question remains the decisive one.

The AI-generated Caitlin Johnstone is very good at explaining why we should distrust the establishment.

My criticism begins one step later:

Why should we trust her?

That is not cynicism. It is the logical consequence of taking her own skepticism seriously.