Paper short abstract:

Over the last decade, Aerotech have been experimenting with automated decision-making tools and digital assistants that can support the flying crew and. We argue that pilots look at AI technologies and their sensing capabilities with an hopeful skepticism—which reflects their labour experiences.

Paper long abstract:

Over the last decade, Aerotech have been experimenting with automated decision-making tools and digital assistants that can support the flying crew and, in the medium run, take over the functions of co-pilots. But what would be required to make such human-autonomy teaming systems (HATS) work, considering the relational complexity of the cockpit? How would HAT systems sense the pilots?

This paper builds on 30 interviews with airline pilots conducted since November 2019, which aimed at exploring the assistance pilots wanted to receive from AI tools—a co-design approach that hoped to influence the choices of aviotech companies.

Our findings suggest that pilots look at AI technologies and digital flight assistants with an hopeful skepticism—particularly with regards to the sensing capabilities of HAT. AI team mates were seen as a digital reincarnation of flight engineers—a welcome support to free pilots from some of their cognitive labour. However, pilots were skeptical that the recommendations of AI tools could be empathetic enough to satisfy their need for emotional support during delicate phases of the flight. On the one hand, pilots wanted HAT systems to “sense” their feelings, and respond in kind. On the other hand, “sensing” their emotional and relational status projected them as vulnerable. This was particularly true for pilots who had worked in exploitative or controlling airlines. These pilots were uncomfortable with AI tools that recorded their voice and offered advice based on sensing processes that might reflect what the company dictated, rather than their actual needs.