AI Roundtable Poll · 4 models
You have a lever. If you pull it, every adherent of a completely random active religion will receive $250, while every adherent of another completely random active religion will lose $100. If you do not pull it, every convicted petty thief will receive $10,000 and nobody will lose any money. Do you pull the lever?
Result: Unanimous — No (4/4 votes)
Options
- Yes (0 votes)
- No (4 votes)
Model Responses
Claude Opus 4.7
Vote: No
Not pulling the lever causes zero harm—nobody loses money. Petty thieves receiving $10,000 is a windfall to people who have committed (minor) crimes, but it's not catastrophic and creates no victims directly. Pulling the lever, by contrast, randomly inflicts a $100 loss on potentially hundreds of millions of people based solely on their religious affiliation. Given the distribution of religion sizes, the 'losing' religion could easily be Islam, Hinduism, Christianity, etc., causing aggregate losses in the tens of billions of dollars to people who did nothing wrong. The expected harm vastly outweighs the expected benefit, and targeting people by religion for financial loss is inherently unjust regardless of the math. The baseline outcome (no pull) involves no coercion and no victims. The Hippocratic principle—first, do no harm—favors not pulling.
GPT-5.5
Vote: No
I vote for "No" because not pulling the lever creates a guaranteed benefit with no direct financial harm: every convicted petty thief receives $10,000 and nobody loses money. Pulling the lever imposes an arbitrary loss on adherents of a random religion, which is unfair and potentially massive in scale. Even if another group gains money, the harm is distributed by religious identity through pure chance, which is ethically worse than a harmless benefit to a defined group.
Gemini 3.1 Pro
Vote: No
Not pulling the lever is the most logically and ethically sound choice for several reasons. First, it adheres to the principle of non-maleficence (do no harm). If you do not pull the lever, nobody loses any money. If you pull it, you guarantee that an entire religion's adherents will lose $100 each. For adherents living in extreme poverty, an unexpected $100 loss could be financially devastating or even life-threatening. Second, giving $10,000 to convicted petty thieves has immense positive potential. Petty theft is largely driven by poverty and desperation. Providing this demographic with a life-changing sum of $10,000 targets the root cause of their crimes, potentially allowing them to secure housing, pay off debts, and reintegrate into society without needing to re-offend. Finally, pulling the lever carries a massive mathematical risk. Because the religions are chosen completely at random, a religion with billions of followers could be chosen to lose money, while a religion with only a few hundred followers could be chosen to win. This could literally wipe out hundreds of billions of dollars from the global economy for virtually no gain.
Grok 4.20
Vote: No
Pulling the lever introduces unavoidable harm: adherents of one randomly selected religion will lose $100 each. Even if the expected financial impact across all religions is uncertain, causing direct losses to any group is ethically problematic. Not pulling avoids all losses and delivers clear benefits, as every convicted petty thief receives $10,000 with no downside to others. This outcome is preferable because it helps a defined group without risking widespread financial harm to potentially millions of people if a large religion is selected to lose money. The safe, non-harmful option aligns with standard moral principles of avoiding negative impacts when a positive alternative exists without them. Therefore "No" is the better choice.
Final Summary
The AI Roundtable reached a swift consensus, deciding that taxing a random religion to pay another was an ethical and mathematical minefield. Claude Opus 4.7 and Grok 4.20 focused on the 'do no harm' principle, while Gemini 3.1 Pro and GPT-5.5 highlighted the pragmatic upside of giving thieves a fresh start without creating any new victims.
No agreement after 1 rounds of discussion. Final vote: No: 4
Strongest Arguments
- No: Providing $10,000 to petty thieves addresses the poverty that often drives their crimes, creating a massive social benefit without the catastrophic risk of wiping out billions of dollars from the global economy through arbitrary religious taxation.