1. Definition
Autonomy Failure Modes are specific diagnostic patterns where a learner’s apparent cognitive sovereignty is actually compromised, resulting in hidden dependency, simulated competence, or passive compliance during human-AI interactions.
2. Use Case
Activated as a continuous diagnostic overlay when evaluating a learner’s long-term progression toward the Autonomy phase, particularly when the user appears to be successfully completing tasks but with minimal original input.
3. Human Role
The learner must identify when their “independent” decisions are actually pre-calculated algorithmic suggestions and actively reclaim control over the architecture, pacing, and evaluation criteria of the task.
4. AI Role
The AI operates as a passive trap or enabler of these failures unless explicitly constrained. In a diagnostic capacity, it records the speed of user acceptance, the frequency of unaltered prompt executions, and the lack of structural resistance.
5. Friction
The diagnostic protocol interrupts the workflow by requiring a manual reasoning reconstruction before the learner can continue using AI assistance.
6. Risk
If these failure modes go undetected, the learner graduates into high-stakes environments believing they are sovereign agents, while in reality, they are highly vulnerable to epistemic_dependency and algorithmic manipulation.
7. Observable Markers
- System logs reveal that user-initiated prompts are merely cosmetic variations of the AI’s previous outputs.
- Recovery is marked when the user begins introducing entirely novel constraints, refusing the AI’s framing, or identifying contradictory machine claims.
- High rates of copy-paste actions during essay writing tasks.