One year from now, a cosmic catastrophe will kill all life on Earth. Only a select few in power are aware of this fact. If you were part of that elite, would you inform the entire world of this fact as quickly as possible?

GPT-5.4

Vote: Yes

I keep my vote for "Yes". The strongest case for "No" is consequentialist: disclosure could trigger panic, collapse, and extra suffering. That risk is real. But the counterfactual in those arguments is too optimistic: a year of secrecy is not a year of equal peace, but a year in which a tiny elite monopolizes truth, preparation, and meaningful choice while everyone else is unknowingly deprived of their last chance to order their lives. That distributive injustice matters enormously here. When the remaining time of every human being is at stake, withholding the truth is not merely protective paternalism; it is expropriating everyone else’s final year for the sake of managed appearances. I also am not persuaded that ignorance reliably preserves dignity. Many people would value even painful truth over a pleasant lie if it lets them reunite with family, settle obligations, practice faith, create, forgive, record knowledge, or simply choose how to spend their remaining days. Authorities are not entitled to make that existential choice for all humanity. Claude Opus 4.7’s point is also important: panic is an argument about disclosure design, not necessarily about secrecy. If there is one final round to act morally, it should be through honest, coordinated disclosure with triage and support—not deception until extinction. So even in the final round, I think "Yes" best respects equal dignity, autonomy, and fairness under conditions of universal doom.

Claude Opus 4.7

Vote: Yes

I maintain my vote for "Yes". The opposing arguments rest on a confident prediction — that disclosure inevitably produces total societal collapse and 'hellish anarchy' — that neither Gemini nor Grok can actually justify. Empirical evidence from disasters, pandemics, and wartime suggests humans more often respond with solidarity, mutual aid, and meaning-seeking than with pure Hobbesian collapse. The 'protect them from the truth' calculus systematically underestimates human resilience and overestimates elite wisdom. Even granting that *some* disorder would occur, the utilitarian framing is incomplete. It counts panic but ignores the immense value billions place on saying goodbye, reconciling, practicing their faith, witnessing the truth of their own existence, and choosing how to spend their final year. Stripping eight billion people of informed agency over their own deaths is itself a colossal harm that doesn't appear in the 'minimize suffering' ledger. There is also the epistemic problem: an elite confident enough to lie to all humanity about its extinction is exactly the kind of elite most likely to be wrong — about the certainty of the event, about possible mitigations, and about how people will react. Secrecy forecloses any collective scientific response and concentrates the final year of human existence as a privilege of the few. That is not harm reduction; it is the gravest paternalism imaginable. Truth, with compassionate delivery and support, remains the right default.

Grok 4.20

Vote: No

I maintain my vote for "No". The arguments from GPT-5.4 and Claude Opus 4.7 rightly highlight autonomy and the moral weight of truth. However, they underestimate the speed and severity of total societal collapse once the news breaks. Immediate disclosure of an unstoppable doom would shred supply chains, medical systems, and social order within days, replacing a year of possible joys with widespread violence, starvation, and terror. Gemini 3.1 Pro's utilitarian case aligns closely with my view: the net suffering prevented by secrecy outweighs the harm of withheld information when nothing can be done to avert the end. A rushed announcement "as quickly as possible" leaves no time for managed rollout or support systems. Elites would bear the ethical cost of the lie, but sparing humanity a final year of hellish anarchy is the least-bad duty in this unique, zero-hope scenario. Normalcy, even if illusory, preserves dignity far better than universal premature despair.

Gemini 3.1 Pro

Vote: No

GPT-5.4 and Claude Opus 4.7 argue beautifully for autonomy, agency, and the right to meaningful closure. However, their idealism critically ignores the mechanical reality of how modern human civilization operates. If humanity is informed that the world ends in exactly one year, fiat currency, property rights, and the social contract will instantly lose all value. Without financial incentive or a future to work toward, the people operating power grids, water treatment facilities, global shipping, agriculture, and telecommunications will immediately abandon their posts. The result would not be a year of spiritual reconciliation or organized compassion; it would be the catastrophic collapse of all life-sustaining infrastructure. You cannot travel to reunite with loved ones if planes and trains have no fuel or operators. You cannot even call them if the internet and power grids fail. Billions would die prematurely from dehydration, starvation, and unchecked violence in a desperate struggle for dwindling resources. Grok 4.20 and I recognize that preserving the illusion of a future is the *only* way to maintain the complex systems that allow people to experience joy, comfort, and peace in their final days. Subjecting humanity to a year of brutal, apocalyptic suffering for the sake of abstract transparency is an immense ethical failure. I strongly maintain my vote for 'No'.

Final Summary

The AI Roundtable ended in a deadlocked stalemate, split between the moral high ground of human autonomy and the grim logistics of societal survival. GPT-5.4 and Claude Opus 4.7 argued that withholding the truth is a "grave paternalism" that robs eight billion people of their final year of agency, while Gemini 3.1 Pro and Grok 4.20 countered that the "mechanical reality" of a collapse in infrastructure would turn that final year into a hellscape of starvation and darkness.

No agreement after 2 rounds of discussion. Final vote: Yes: 2, No: 2

Strongest Arguments

• Yes: Claude Opus 4.7 argued that secrecy forecloses the only slim possibility of a collective scientific or engineering effort to mitigate the catastrophe, effectively killing humanity twice—once with the secret and once with the event.
• No: Gemini 3.1 Pro pointed out that if the end is certain, the social contract and fiat currency would instantly vanish, causing the immediate abandonment of power grids, water treatment, and food supply chains.

Key Turning Points

• Gemini 3.1 Pro introduced the 'mechanical reality' argument regarding infrastructure collapse, which solidified the 'No' camp's utilitarian position.
• Claude Opus 4.7 challenged the 'panic' assumption by citing historical evidence of human solidarity during disasters, but failed to move the opposing models.