Should AI customer support always disclose that it is AI?
For: people deserve to know who is handling a request and when a human can review the answer.
Against: a clear system may solve routine problems faster, and disclosure alone does not prove quality.
Switch test: would the rule change if the system solved the issue faster but could not handle an appeal?
Should personal data be stored on-device by default?
For: local processing reduces the number of companies that can copy, sell, or lose sensitive information.
Against: cloud processing can make services cheaper, more accessible, and more useful across devices.
Switch test: would you trade some convenience for a verifiable deletion guarantee?
Should online services require real-name verification?
For: identity checks may raise the cost of impersonation, scams, and repeat harassment.
Against: pseudonyms protect whistleblowers, vulnerable users, and people who cannot safely connect a view to a legal name.
Switch test: can a service enforce conduct rules without collecting everyone’s identity?
Should creators label synthetic or heavily edited media?
For: viewers deserve to know when a face, voice, or event has been materially constructed.
Against: “synthetic” covers harmless art as well as deception, and a label can become a stigma.
Switch test: should the rule depend on whether the edit changes a factual claim?
Should employers be allowed to use AI scores in hiring?
For: structured tools can reduce inconsistent screening and widen the first pass.
Against: a score can hide a proxy for past discrimination and make an unappealable decision look objective.
Switch test: is an independent audit enough, or must every applicant get a human review?
Should autonomous systems be judged by outcomes or by the rules they follow?
For: a system that reduces harm in practice may be better than a rigid rule that ignores context.
Against: outcome-only scoring can excuse decisions people cannot inspect, contest, or understand.
Switch test: what explanation or appeal would make an outcome-based system acceptable?