ScenarioSuperintelligence control and containment assurance

Superintelligence Control Assurance Architect

Also called SI Control Assurance Architect

A Superintelligence Control Assurance Architect would judge whether claims of control over broadly superhuman systems hold up. It depends on how AI develops, and matters only if such systems emerge and institutions want credible oversight.

What they do

The architect would take a control claim for a highly capable system and try to break it. They would map what the claim assumes about the system, the operators and the environment. They would design adversarial tests that do not rely on humans out-thinking the system. They would check where formal methods apply and where they stop. They would work with security, safety and governance teams, and with outside reviewers. They would write down what is known, what is assumed and what cannot be tested. They would also say when evidence is too weak to support a claim.

What a strong hire delivers

In the first months, a strong hire would produce a conditional control assurance case that names its unknowns plainly. They would agree with leadership what counts as independent evidence and how claims could be shown false. They would set out the limits of control that the company would disclose. They would also leave a review process that outside parties could inspect and challenge.

How you can tell they are good

Ask them to take a control claim and explain how it could fail. Strong candidates separate evidence from argument and say what would change their mind. They are comfortable saying a claim cannot be verified. Ask for past work where they found a flaw in a safety or security case. Be wary of anyone offering certainty or a tidy toolkit.

Why it is hard to fill

If this role emerged, the people would be scarce because no one has done it. Titles would vary and the work sits across control research, security, verification and system safety. Few people combine deep technical skill with the discipline to write honest limits. Many candidates would be strong in one area and shallow in the rest. Any claim of direct experience should be treated with caution.

Where to find them

Today they would come from AI control research, security engineering, formal verification and safety-critical system assurance. Look in frontier lab safety teams, independent evaluation groups and AI governance bodies. Red teams, high-assurance software groups and system safety engineers in regulated industries also produce people with the right habits. Scenario analysis experience would help.

Interview questions for a Superintelligence Control Assurance Architect

  1. Tell us about a time you found a serious flaw in someone else's safety or security claim. How did you show it?

    A strong answer gives a specific case, the evidence used and how the owner responded. It shows independence and clear reporting. Worry if the story is vague, or if the flaw was found only by luck.

  2. What would make you say a control claim for a system far more capable than its overseers is not credible?

    Good answers name concrete failure points: untestable assumptions, reliance on the system's own reports, and weak independence. They stay specific. Worry if they simply assume capable overseers or trust the system's cooperation.

4 more questions, with what a strong answer covers, are in the Role Kit.

Job description you can post

We are building the ability to judge whether claims of control over very capable AI systems are credible. We would like an architect who can test those claims adversarially, state what is unknown and report findings plainly, including unwelcome ones.

What you will do

  • Assess control claims for highly capable systems and identify the assumptions behind them
  • Design adversarial tests and evidence standards that do not depend on human-level oversight
  • Write conditional assurance cases with explicit unknowns
  • Define how claims could be shown false
  • Work with security, safety and governance teams
  • Support independent external review
  • Report the limits of control clearly to leadership

What you will bring

  • Demonstrated work finding flaws in safety, security or control arguments
  • Strong grounding in AI control, security, formal verification or system safety
  • Clear written reasoning about uncertainty
  • Experience designing adversarial evaluations or red team exercises
  • Ability to explain technical limits to non-specialists
  • Independent judgement under disagreement

Nice to have

  • Experience with formal verification of complex systems
  • Background in AI governance or assurance
  • Scenario analysis for frontier AI

Free to use and adapt. If it saves you time, a link back to this page is appreciated.

Depends on a scenario

Superintelligent systems with major strategic capability

What AI would need to do
Systems substantially exceed human expert groups across broad consequential cognitive tasks.
What organisations would need to decide
Such systems are developed or deployed and institutions seek credible control and external oversight.
What would rule it out
No broad superhuman capability or no usable, governable deployment emerges.

We give no date. The role becomes real if both conditions hold.

Skills that matter

  • AI controlAdvancedLimit harm from potentially untrusted AI behaviour
  • Formal verificationAdvancedProve stated properties under explicit assumptions
  • AI governance and assuranceWorkingAssign accountability and verify lifecycle controls
  • Frontier scenario analysisWorkingExplore conditional outcomes without claiming predictions

Other roles in superintelligence control and containment assurance

ScenarioSuperintelligent System Containment ResearcherA containment researcher would test whether enforceable limits can hold against broadly superhuman AI. The role depends on how AI develops, and it…ScenarioSuperintelligence Intervention Protocol DesignerA Superintelligence Intervention Protocol Designer would specify who may intervene in a broadly superhuman system, how, and whether it would work…ScenarioSuperintelligence Strategic Risk AnalystA Superintelligence Strategic Risk Analyst would map how broadly superhuman systems could affect institutions and security. Whether the job exists…

Related roles in other families

ScenarioGeneral-Purpose Autonomy Assurance ArchitectGeneral-purpose autonomy assuranceA General-Purpose Autonomy Assurance Architect would design how an organisation gains justified confidence in systems given broad, consequential…ScenarioCross-Domain Capability Certification DesignerGeneral-purpose autonomy assuranceA Cross-Domain Capability Certification Designer would define what a general-purpose AI system can credibly be said to do across consequential…ScenarioAutonomous Delegation Accountability OfficerGeneral-purpose autonomy assuranceAn Autonomous Delegation Accountability Officer would keep track of who authorised what when AI systems carry out broad work. The role depends on…ScenarioGeneral-Purpose System Decommissioning PlannerGeneral-purpose autonomy assuranceA General-Purpose System Decommissioning Planner would prepare the controlled withdrawal of autonomous systems that run deeply inside an…

How we wrote this

Role definitions are our own editorial work, checked against occupational classifications, standards, research and employer pages. Titles vary by company, and this is a description of the work, not a job advert. Scenario roles describe responsibilities that depend on stated conditions; they are not forecasts.

← Superintelligence control and containment assuranceAll new job families

Start a search

Hiring a Superintelligence Control Assurance Architect?

Tell us the work. We find the people already doing it, whatever their job title says, and you meet them already interviewed for your role.