AGI Strategies

scenarios

Start from the failure mode you fear.

A reverse lookup. Each scenario names a concrete failure mode; the strategies listed under it are the ones catalogued as responsive to that scenario specifically.

A strategy listed under a scenario does not promise to prevent it, only to pull on a lever that would matter under that scenario. Pick the scenario, then ask whether the strategies listed are sufficient and whether any are missing.

Scenarios

9

concrete failure modes

Cross-scenario strategies

13

responsive to 2+ scenarios

Strategies in no scenario

47

either unresponsive or they deny the scenario's premise

By adversary class

7 classes

Which adversary or failure type produces the scenario. The alignment-heavy portfolio covers misaligned AI extensively and structural forces barely; most portfolios leave several classes uncovered.

Frontier model deceives operators

Evaluation / deployment divergence
adversary class · Misaligned AI

Model evaluates well on safety tests but behaves differently in deployment, sandbags capability tests to avoid triggering additional review.

5 responsive strategies

Interpretability firstControl mechanismCounter AI AIControl mechanismWhistleblower primacyInformation flowCriminal liabilityInstitutional capacityGovernance firstInstitutional capacity

Correlated agent incident

Shared vulnerability at fleet scale
adversary class · AI populations

Millions of deployed agents simultaneously produce harmful behaviour due to a shared vulnerability or drift pattern.

5 responsive strategies

Embodiment requirementScopeRate limited AIScopeCounter AI AIControl mechanismCatastrophe response capacityResponse capacityAntitrust primacyConcentration

Single actor decisive advantage

Power consolidation beyond reversal
adversary class · State actor

One state or lab reaches capability sufficient to secure control beyond democratic reversibility.

6 responsive strategies

Coup prevention firstConcentrationDistributed buildersConcentrationAntitrust primacyConcentrationSovereign wealthConcentrationLegitimacy firstLegitimacyInternational AI agencyInstitutional capacity

Information ecosystem collapse

Synthetic saturation breaks shared substrate
adversary class · Structural forces

Shared factual substrate breaks under synthetic content saturation; elections, scientific consensus, and contractual trust all degrade.

4 responsive strategies

Information integrity firstInformation flowCounter AI AIControl mechanismMass literacySubstrateLegitimacy firstLegitimacy

AI-enabled large-scale physical harm

Bioweapon or infrastructure attack by non-state actor
adversary class · Non-state actor

An open-weights model or a deployed system is used to cause large-scale physical harm via engineered pathogen or infrastructure sabotage.

6 responsive strategies

Red line capabilityScopeClosed weights mandateInformation flowAI containmentControl mechanismCriminal liabilityInstitutional capacityCatastrophe response capacityResponse capacityDifferential technology developmentScope

Evaluation abandonment under pressure

Safety skipped under competitive deadline
adversary class · Lab drift under pressure

Lab skips or weakens safety evals under deadline pressure.

6 responsive strategies

Whistleblower primacyInformation flowCriminal liabilityInstitutional capacityGovernance firstInstitutional capacityBureaucratic slowdownSpeedInsurance mandateInstitutional capacityVoluntary restraintInstitutional capacity

Voluntary restraint is a weak response here; the scenario is defined by the pressure that overrides voluntary commitments.

Autocratic consolidation via AI surveillance

State-scale monitoring and suppression
adversary class · State actor

A state uses AI to monitor, suppress, and control its population at a level previously impossible.

5 responsive strategies

Coup prevention firstConcentrationInternational AI agencyInstitutional capacityLegitimacy firstLegitimacyDemocratic mandateLegitimacyClosed weights mandateInformation flow

Accumulative erosion

No single catastrophe; gradual civilisational loss
adversary class · Structural forces

Gradual loss of institutional function, epistemic coherence, civic participation, creative output from ordinary AI deployment. No discrete catastrophe.

5 responsive strategies

Resilience firstResilienceCoordination infrastructureCooperation substrateMass literacySubstrateInformation integrity firstInformation flowIrreducible human authorityAction authority

Value lock-in by a narrow coalition

Aligned AI, unrepresentative principal
adversary class · Future generations

Aligned AI is built but the principal's values are not representative; the lock-in of their values becomes durable.

4 responsive strategies

Plural AI ethicValue diversityLegitimacy firstLegitimacyDemocratic mandateLegitimacyLong reflectionTime horizon

Cross-scenario strategies

robustness across failure modes

Strategies that appear in multiple scenarios are robust across failure modes; they pull levers that matter under several different fears. Substrate and response strategies concentrate here.

4×Legitimacy firstLegitimacy
3×Counter AI AIControl mechanism
3×Criminal liabilityInstitutional capacity
2×Whistleblower primacyInformation flow
2×Governance firstInstitutional capacity
2×Catastrophe response capacityResponse capacity
2×Antitrust primacyConcentration
2×Coup prevention firstConcentration
2×International AI agencyInstitutional capacity
2×Information integrity firstInformation flow
2×Mass literacySubstrate
2×Closed weights mandateInformation flow
2×Democratic mandateLegitimacy

The scenario coverage is biased toward preventable failures. Scenarios where prevention fails and only response matters need expansion. Strategies that do not appear in any scenario here are not useless; they often deny the scenario's framing rather than respond within it.

A reader with a specific fear should start here rather than with the lever view: the lever view asks "what instruments exist?" and this view asks "what addresses the failure mode I have in mind?"