AI is machine-based technology that, for human-defined explicit or implicit objectives, infers from the machine and human-based input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence real/physical or vir- tual environments. The use of AI can be divided between implementations of fixed and deterministic algorithms, or implementations that continue to change through learning in the operational en- vironment. Hence, in the interest of concise- ness and facilitating interpretation by intended users with varying expertise, AI will be defined as a machine-based system designed to mimic the neural Pilot interventions are actions taken to mitigate a threat or undesired aircraft state and maintain safety margins by interrupting/overriding the performance of a task being ex- ecuted or ignored by automation dur- ing normal, abnormal, or emergency conditions or otherwise when en- countering an uncertain environment and/or unforeseeable situations. That definition is composed of ter- minology related to aircraft and per- sonnel safety together with the overall environmental situation, all of which need to be defined. • Emergency condition. Safety of the aircraft or of people on board or on the ground is endangered for any reason, and it is no longer pos- sible to continue the flight. • Abnormal condition. Safety of the aircraft or persons on board or on the ground is not in danger, but it is no longer possible to continue the flight using normal procedures. • Normal condition. Safety of the aircraft or persons on board or on the ground is not in danger and it is possible to continue the flight using normal procedures. networks of a human pilot . Pilot Intervention • Uncertain environment. A lack of information about the environment that results in uncertainty about the outcome (i.e., consequence) as well as uncertainty about the
probability of a certain situation occurring in that environment. • Unforeseeable situation. Con- ditions that could not reason- ably have been anticipated or known beforehand. • Threat. An event, condition, or error that requires attention and management to maintain safety margins. • Automation. The independent ac- complishment of a function or task by a device or system that was for- merly carried out by a human. The theoretical basis for a pro- posed PIRS is grounded in an adap- tation of the Turing Test originally postulated by the British philosopher, mathematician, and cryptanalyst Alan Turing as a method for determining Pilot Intervention Rating Scale (PIRS)
whether a computer could mimic a human’s response. The experiment involved asking a human interrogator to distinguish between two remotely located respondents—one human and one computer. The premise was that if the interrogator could not distinguish the computer from the human, the computer would be con- sidered to have successfully mimicked the human. The Turing Test concept along with the rationale and methodology from the various existing pilot rating scales was adapted for this proposed PIRS. Considerations included use of an objective quantitative multilevel score that could easily be interpreted by those with limited expertise and that categorized the findings for comparisons and/or consolidation with the findings from other flight tests of AI aircraft systems.
Figure 3. Pilot Intervention Rating Scale
Pilot Action PIR
Consequence
Ad Hoc Emergency Procedure Existing Emergency Procedure Ad Hoc Abnormal Procedure Existing Abnormal Procedure Ad Hoc Normal Procedure Existing Normal Procedure
Avoided Major Aircraft Damage and/or Serious Injury or Fatalities
6
Emergency Condition
5
Yes
LEVEL 3
Avoided Minor Damage to Equipment or Property and/or Minor Injury
4
Abnormal Condition
Safety of Aircraft/Person?
No
3
LEVEL 2
No
Avoided Inconvenience with No Safety Consequences
2
Continued Flight Using Normal Procedures?
Normal Condition
Yes
1
LEVEL 1
Yes
No Inconvenience, or Safety Consequences
Normal Condition
No Pilot Action
Pilot Intervened?
2
No
Start Here
Source. Author
26 | DEFENSE ACQUISITION | November-December 2025
Made with FlippingBook - Online Brochure Maker