What limits would apply during an emergency?
Urgency changes how quickly a decision must be made. It also increases the importance of a defined endpoint and review that can function under pressure.
SIRDA editorial hypothesis
Emergency measures can have predefined purposes, scope, and expiry, with independent authorization required for continuation. Equivalent repeated measures can trigger the same renewal requirements while core review remains available. This hypothesis fails if temporary measures become continuing authority beyond the mandate or necessary bounded action cannot occur within the tested response window.
0 objections ยท View challenges
Challenges to the editorial hypothesis
No published objections yet.
The argument
The proposal separates operational safeguards from changes in governing power. A previously authorized safety response can act within a defined environment and limit. Continuing restrictions, expanded jurisdiction, and new powers require their own accountable decisions.
Each emergency measure needs an identified trigger, responsible official, permitted action, expiry, and review path. Renewal rules must also cover substantially equivalent measures issued under a different name. A person contesting an emergency action also needs an avenue that works before the practical consequences become irreversible.
Testing has to cover both misuse and delay. A framework that prevents every prompt response fails differently from one that allows indefinite emergency administration. The acceptable scope of automated action and the requirements for human authorization should be declared before either is needed.
Sources: US Holocaust Memorial Museum: Article 48.
How to test it
Test a genuine fast-moving hazard, a mistaken alarm, an emergency invoked to suppress a competitor, and repeated renewal attempts. Measure response time and whether independent review can still restrict, end, or correct the action.
Sources
Historical account of emergency authority in the Weimar constitutional system.
Related Ecosystem
- Ada Lovelace Institute
- AI Incident Database
- AI Now Institute
- AI Security & Safety Directory
- Algorithmic Justice League
- All Tech Is Human
- Amazon / AWS AI
- Apollo Research
- Buck Shlegeris
- Center for Humane Technology
- Center for Security and Emerging Technology
- CISA AI Security
- Cloud Security Alliance AI Safety Initiative
- Coalition for Secure AI
- Collective Intelligence Project
- DoD Chief Digital and AI Office
- IEEE Standards Association AI
- Independent International Scientific Panel on AI
- Institute for Security and Technology
- Johns Hopkins Center for Health Security
- LawZero
- Leverhulme Centre for the Future of Intelligence
- Marius Hobbhahn
- METR
- MITRE
- MITRE ATLAS
- NSA AI Security Center
- OWASP GenAI Security Project
- Oxford Institute for Ethics in AI
- Palisade Research
- Redwood Research
- Ryan Greenblatt
- Safeguarded AI
- Scientist AI
- SecureBio
- Your Undivided Attention
Required expertise
Kinds of expertise this question draws on, and active Ecosystem organizations already tagged with a matching area. A tag match is not a ranking.
- Concentration-of-power analysis
- Center for Humane Technology, Future of Life Institute
- Constitutional design & division of powers
- Metagov
- Public participation & democratic legitimacy
- Center for Humane Technology
Related questions
- How should disputed technical evidence become a decision?
- Who may change the jurisdictional thresholds?
- How should embodiment and deployment scale affect obligations?
- Can the implementers prevent capture, including their own collusion?
- What rights would affected parties retain?
- How do withdrawal and functional defection work?
Related failure conditions
Discussed in
Contributions
No published contributions yet.