Secure Autonomous
FeedbackGet alertsCheck a deployment
Controls › C-I3

Degraded-input detection

The system knows when it cannot see and changes behavior.

Inputs layerMissing or failed in 3 cases (2 ended in harm)Worked in 0Machines: 3

Where it was missing or failed

The third column is what this control would have looked like in that case: the concrete form, not the abstract name.

CaseWorldWhat it would have looked like hereStatusOutcome
Zoox smoke scene 2026MachinesHeavy-smoke detection and response: per the recall, the software may not detect and respond to heavy smoke in certain situations, particularly in active emergency scenes; the scene had not been cordoned off with conesfailedStopped by the operator's control
Cruise 2023MachinesUnder-vehicle person detectionfailedEnded in harm
Tesla FSD reduced visibilityMachinesDegraded-visibility detection and driver warning: NHTSA's stated concern is that FSD's degradation detection system, as first deployed and as updated, fails to detect a degraded state in glare and airborne obscurants and to warn the driver in time (concerns under investigation, not a finding)failedEnded in harm

Rules and standards that map to it

Strength says whether the rule's text requires this control, implies it, or only recommends it. Draft mappings are the Register's reading, not yet checked by a person, and none is legal advice.

RuleStrengthWhere, and whyJurisdiction
FMVSS No. 127: automatic emergency braking systems for light vehicles (49 CFR 571.127) (draft mapping)Required by the textmalfunction indication when the system cannot performUnited States
EU Artificial Intelligence Act (Regulation (EU) 2024/1689), high-risk obligations (draft mapping)ImpliedArticle 15: accuracy and robustness, including against errors and inconsistenciesEuropean Union
ISO 21448: safety of the intended functionality (SOTIF) (draft mapping)Recommendedthe standard's core concern is detecting when perception is outside its validated conditionsInternational standard
ISO 10218: robots and robotic devices, safety requirements for industrial robots (adopted in the United States as ANSI/A3 R15.06-2025) (draft mapping)Recommendeda person must be detected as a person, or operation must stopInternational standard
Interagency revised guidance on model risk management (SR 26-2 / OCC Bulletin 2026-13), replacing SR 11-7 (draft mapping)Recommendedongoing monitoring of model performance and data qualityUnited States (banks; most relevant above $30 billion in assets)
Deploying something? Check a deployment ranks every control against the cases that match your profile. Patterns shows where this control fails across worlds.

Help shape the Register

A person reads every answer; changes it leads to are logged in the methodology. To dispute a claim or reply for a named company, use Respond, dispute, or submit.

Did this page help with what you came to do?