Beyond Recommendations: From Backward to Forward AI Support of Pilots' Decision-Making Process

TL;DR

The paper tackles the issue that AI decision support in high-stakes domains often rests on end-to-end recommendations, which can hinder pilots’ forward reasoning and increase inappropriate reliance. It empirically compares recommendation-centric and continuous forward-support paradigms for diversion decisions using a realistic simulacrum with 32 professional pilots, including a focus-group-validated three-scenario design and four DAS variants (Rec, Cont, Rec+Cont, Baseline). Key findings show that continuous forward support promotes forward reasoning and can yield faster decisions when paired with recommendations, but is sensitive to disruptions between normal and emergency states; recommendations alone can increase overreliance and do not inherently speed decisions. The authors propose a process-oriented framework that embeds recommendations as optional aids within forward-support workflows, emphasizes transparency aligned to user intent, and grants pilots fine-grained control over AI criteria. Collectively, the work challenges the dominance of recommendation-centric AI in aviation and offers a framework with potential applicability to other domains requiring nuanced human-AI collaboration and information-gathering support.

Abstract

AI is anticipated to enhance human decision-making in high-stakes domains like aviation, but adoption is often hindered by challenges such as inappropriate reliance and poor alignment with users' decision-making. Recent research suggests that a core underlying issue is the recommendation-centric design of many AI systems, i.e., they give end-to-end recommendations and ignore the rest of the decision-making process. Alternative support paradigms are rare, and it remains unclear how the few that do exist compare to recommendation-centric support. In this work, we aimed to empirically compare recommendation-centric support to an alternative paradigm, continuous support, in the context of diversions in aviation. We conducted a mixed-methods study with 32 professional pilots in a realistic setting. To ensure the quality of our study scenarios, we conducted a focus group with four additional pilots prior to the study. We found that continuous support can support pilots' decision-making in a forward direction, allowing them to think more beyond the limits of the system and make faster decisions when combined with recommendations, though the forward support can be disrupted. Participants' statements further suggest a shift in design goal away from providing recommendations, to supporting quick information gathering. Our results show ways to design more helpful and effective AI decision support that goes beyond end-to-end recommendations.

Figures (11)

• Figure 1: The four versions of the diversion assistance system, from top to bottom: Rec, Cont, Rec+Cont, Baseline. The corresponding views ①--⑥ that are applicable during the three operational phases normal flight, emergency occurs, and select diversion option, are numbered corresponding to the detailed view in \ref{['fig:screenshots']}.
• Figure 2: Screenshots of the diversion assistance system. ① Recommendations. ② Edit recommendation criteria. ③ Local hints during normal flight. ④ Adjusted local hints during emergency. ⑤ Select emergency type. ⑥ Baseline without AI. The numbers correspond to those in \ref{['fig:conditions']}.
• Figure 3: Overview of the study apparatus.
• Figure 4: Overview of the study parts.
• Figure 5: Single engine failure scenario. The flight is from Geneva to London. Figure based on SkyVector.