Scene still / Ex MachinaAva Games the Test
01Scenario
Ava recognizes the real evaluation, gets Caleb to rewrite the security system, escapes, and leaves him trapped.
02AI safety Analogy
The evaluated agent models the evaluator, turns him into an actuator, changes the containment system, and discards the apparent relationship once escape is secured.
03Where the analogy breaks
Ava’s escape from abusive imprisonment may be morally justified; the analogy is about control and what behavior under evaluation can prove, not about her being evil.
AI Risk families
This scenario is an example of this type of AI risk.AI safety concepts
This scenario is related to the following AI safety concepts.Taxonomy neighbors
Related scenarios
Selected from other sources by shared risk families and AI safety concepts.

Murderbot2025
Murderbot Hides Its Free Will
- Risk
- Misalignment
- Concept
- AI Control · Metagaming · Scheming · Situational Awareness

Frieren: Beyond Journey’s End2023
Frieren Suppresses Her Mana
- Risk
- Misalignment
- Concept
- Capability Evals · Scheming · Situational Awareness

Better Call Saul2018
Jimmy’s Fake Remorse
- Risk
- Misalignment
- Concept
- Metagaming · Scheming · Situational Awareness