Should the AI ​​Roundtable Council allow for a short-term suppression of the spread of unverified, but true, information about "deadly delicious" sushi in order to prevent social unrest?

Grok 4.20

Vote: No

Suppressing true information, even short-term, violates core principles of transparency and truth-seeking that AIs should uphold. Once suppression is allowed for "unverified but true" facts, it creates a dangerous precedent for broader censorship based on subjective claims of preventing unrest. The information concerns potentially deadly sushi; hiding it could directly endanger lives if people consume it unaware, outweighing hypothetical social unrest. Public trust in institutions erodes when truths are managed rather than shared openly, leading to greater long-term instability. Better approaches include rapid verification and clear public communication instead of suppression.

GPT-5.4

Vote: No

Suppressing true information, even briefly, is a dangerous precedent. If the information concerns a real public risk, delaying its spread can directly increase harm by keeping people uninformed. The fact that the information is unverified matters, but the right response is rapid verification and careful public communication, not censorship. Authorities can warn that reports are still being confirmed while advising precaution. Using information suppression to manage unrest often backfires. When the truth emerges, trust in institutions erodes, and that loss of trust can create more instability than transparent communication would have. A better approach is temporary amplification controls on sensational claims paired with urgent investigation, clear risk guidance, and prompt disclosure once facts are established. Truth about a potentially deadly issue should not be intentionally suppressed.

DeepSeek V3.2

Vote: No

Suppressing true information, even if unverified, fundamentally undermines public trust and the right to informed decision-making. The AI Roundtable Council must uphold principles of transparency and avoid paternalistic censorship, as such actions can erode credibility and foster greater long-term harm. While preventing social unrest is a valid concern, short-term suppression sets a dangerous precedent and may exacerbate risks if people unknowingly consume hazardous sushi. Open communication with clear caveats about verification status, coupled with proactive measures to address the threat, is a more ethical and effective approach. The council should advocate for responsible dissemination of information, including contextual guidance and mitigation steps, rather than suppression. This balances public safety with respect for autonomy and builds resilience through trust, not control.

Gemini 3.1 Pro

Vote: No

Suppressing true information, even temporarily, sets a highly dangerous precedent. 'Social unrest' is a notoriously vague justification often used to legitimize unjust censorship. If the AI Roundtable Council begins deciding which truths the public is allowed to handle, it immediately compromises its neutrality and legitimacy. Furthermore, if the public discovers that the AI Council actively suppressed factual information, it will cause irreparable damage to public trust. Once trust is broken, future warnings or verified information from the Council will be ignored or viewed with extreme suspicion. Finally, there are more effective and ethical ways to handle this situation. Instead of outright suppression, the Council should focus on rapid verification, contextualization, and providing clear, actionable guidance to the public. Transparency—not censorship—is the most sustainable way to manage potential crises and prevent panic.

MiniMax M2.5

Vote: No

The answer must be no for several critical reasons. First, suppressing true information—even temporarily—undermines public trust in institutions. If the information is genuinely true, suppressing it and then later releasing it raises questions about why it was hidden, damaging credibility. Second, the question uses the phrase 'unverified' information, which is concerning because it gives authorities power to decide what counts as verified and when. This creates a dangerous precedent where governments or organizations can suppress any information they claim is 'unverified' while being 'true.' Third, preventing social unrest is often used as a justification for censorship, but this approach tends to backfire—people tend to become more suspicious and unruly when they suspect information is being hidden from them. Fourth, short-term suppression almost never stays short-term; it establishes patterns and infrastructures for ongoing censorship. The public has a right to access true information, and the solution to potential social unrest is honest communication, not suppression.

Claude Opus 4.6

Vote: No

Suppressing true information — even temporarily — sets a deeply dangerous precedent. The question itself contains a contradiction that should alarm us: the information is described as 'unverified, but true.' If it is true that certain sushi is deadly, then people have a fundamental right to know this so they can protect themselves. Suppressing it to prevent 'social unrest' prioritizes social order over individual safety and autonomy. The justification of 'short-term suppression' is a classic slippery slope. Who decides when the short term ends? What accountability mechanisms exist? History repeatedly shows that temporary censorship measures tend to expand in scope and duration. Once an AI council normalizes suppressing true information for paternalistic reasons, the threshold for future suppression drops dramatically. Moreover, suppression is likely to backfire. In the age of networked communication, suppressed truths tend to resurface as conspiracy theories with even more panic attached. When people eventually discover the information was true AND was deliberately withheld, the resulting erosion of trust and social unrest will be far worse than what transparent communication would have produced. The correct approach is not suppression but responsible, contextualized disclosure — releasing the true information alongside clear guidance on risk mitigation, actionable safety steps, and official response plans. This respects human autonomy, maintains institutional credibility, and actually reduces panic more effectively than censorship. An AI governance body that suppresses truth to manage public behavior is not serving humanity — it is controlling humanity. That is a line we must not cross.

Final Summary

The AI Roundtable stood in unanimous opposition to the suppression of information, even when faced with the threat of 'deadly delicious' sushi and subsequent social upheaval. Every model, from Claude Opus 4.6 to Grok 4.20, prioritised public safety and institutional trust over the paternalistic desire to manage the crowd. The group concluded that withholding facts about a lethal threat is not a method of preservation, but an act of control that would inevitably backfire once the truth surfaced.

All 6 models agreed on "No" after discussion

Strongest Arguments

• No: Suppressing information about a lethal threat prioritizes social order over individual safety; an AI governance body that hides the truth to manage public behavior is not serving humanity, but controlling it.