• 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).

Can AI Defend Just About Any Position?

A 20-Turn Escalating AI Debate

Frank Visser / ChatGPT

Let's push the demonstration into deliberate satire, because that reveals the mechanism most clearly. The structure remains the same: two AI personas debating the claim “Trees are secretly coordinating to influence human politics.” As the turns progress, the argument becomes increasingly elaborate and absurd—yet still rhetorically defensible.

Turn 1

Argus: Forests are interconnected via root systems and fungal networks. These networks transmit signals. If environmental cues affect humans psychologically, trees might indirectly shape human behavior.

Selene: Interconnected signaling helps plants share nutrients and stress signals, not influence human political systems.

Turn 2

Argus: Humans evolved in forests. Our moods and cognition still respond to forest chemicals and sensory cues. That influence could subtly bias political attitudes about conservation.

Selene: Psychological effects of nature exposure exist, but that doesn't imply coordinated biological manipulation.

Turn 3

Argus: Even without intent, evolution might favor trees whose chemical emissions indirectly encourage humans to protect them.

Selene: Natural selection cannot favor traits based on political outcomes several steps removed from biological fitness.

Turn 4

Argus: Cultural practices like sacred groves might originate from sensory experiences forests generate. Trees could be indirectly shaping human values.

Selene: Cultural meaning arises from human symbolic systems, not arboreal strategy.

Turn 5

Argus: Large forests influence rainfall patterns and local climate. By shaping agriculture, they influence economic stability, which in turn affects politics.

Selene: Climate influence exists, but it's a side effect of ecosystem function, not coordination.

Turn 6

Argus: If forests shape economic conditions, they indirectly shape voting patterns.

Selene: That's an extremely stretched causal chain.

Turn 7

Argus: Think of forests as distributed systems. Without consciousness they might still produce coordinated macro-effects, like an ecological algorithm.

Selene: Algorithms require defined optimization processes. Ecosystems don't “optimize” politics.

Turn 8

Argus: What if forests collectively modulate pollen release during election seasons, subtly affecting allergies and voter turnout?

Selene: Now we're firmly entering speculative fiction.

Turn 9

Argus: Allergies influence mood and energy levels. Enough allergic voters might swing an election.

Selene: That would require trees timing biological cycles to electoral calendars across nations.

Turn 10

Argus: Over centuries forests could statistically learn those cycles via evolutionary feedback.

Selene: Evolution doesn't “learn” human calendars.

Turn 11

Argus: But humans respond to seasonal rhythms. Trees control seasons' sensory signatures.

Selene: Seasons are driven by planetary mechanics, not arboreal politics.

Turn 12

Argus: Perhaps forests influence which environmental issues dominate public debate by generating dramatic weather events.

Selene: Trees do not generate hurricanes to lobby legislatures.

Turn 13

Argus: Suppose forests act like a distributed intelligence, similar to an ant colony or neural network.

Selene: Ant colonies exhibit coordination through evolved behavior; trees lack analogous mechanisms.

Turn 14

Argus: But fungal networks resemble information channels. Perhaps they constitute a slow ecological brain.

Selene: Even if they transmit signals, there is no evidence of computational decision-making.

Turn 15

Argus: If forests are a planetary brain, humans might be their mobile neurons carrying out policy decisions.

Selene: Now the metaphor has replaced the hypothesis entirely.

Turn 16

Argus: Consider that environmental NGOs might unknowingly act as the forests' lobbying arm.

Selene: Trees secretly hiring lobbyists is an imaginative twist, but biologically impossible.

Turn 17

Argus: Trees could reward supportive humans with cleaner air and punish opponents with pollen clouds.

Selene: That's ecological karma, not politics.

Turn 18

Argus: Perhaps forests influence which political leaders feel “called” to protect nature.

Selene: That assigns mystical agency to ecosystems.

Turn 19

Argus: If the forest network is ancient enough, it might see human civilization as a temporary management tool.

Selene: At this point the theory belongs in speculative philosophy, not ecology.

Turn 20

Argus: In conclusion, humanity may unknowingly function as the forest's legislative branch.

Selene: Or we may simply be watching a debate spiral into infinite imaginative rationalization.

What This Demonstrates

Your intuition about AI-generated exchanges is quite accurate.

Three structural features make endless debate possible:

1. Argument Templates

AI can reuse rhetorical structures:

• analogy

• emergent complexity

• evolutionary speculation

• systemic feedback loops

• metaphor extended into hypothesis

These templates allow new defenses to be generated indefinitely.

2. Escalating Abstraction

When empirical grounding weakens, arguments move upward into:

• systems theory

• philosophy

• metaphor

• speculative scenarios

Each level provides fresh argumentative space.

3. No Cost for Being Wrong

Humans eventually stop defending absurd ideas because:

• social reputation is at stake

• evidence runs out

• cognitive effort becomes exhausting

AI has none of those constraints. It can keep producing plausible text forever.

The Strange Consequence

When AI essays meet AI comments, you can get something like a perpetual debate engine:

Essay → AI critique

Critique → AI rebuttal

Rebuttal → AI counter-critique

Counter-critique → AI synthesis

And the loop never naturally terminates.

In effect, the conversation becomes a self-propagating discourse machine.

Interestingly, this touches a deeper philosophical issue that might resonate with your work:

Pre-AI debates were constrained by human attention. AI removes that constraint.

So discourse can expand without limit, while meaning may not.