How High Can AAM Go with AI?

EASA’s EUROCONTROL shows us 6 ways Human-AI teaming could take flight.

October 27, 2024 | Dave Clarke

The potential for artificial intelligence (AI) to board tomorrow’s aircraft, partner with humans, and propel the aviation industry to new heights is a tantalizing proposition. Aviation + AI has the potential to reduce costly delays and shrink the industry’s carbon footprint while enhancing safety in the cockpit, facilitating single pilot operations, and managing what promises to be very densely occupied airspace as drone operations above urban areas begin to soar.

With that in mind, the European Union Aviation Safety Agency (EASA) is suggesting six categories of Human-AI teaming ranging from machine learning support to fully autonomous AI. Key to understanding this man and machine partnership is something known as Human Factors. A multidisciplinary field, Human Factors studies how people interact with products, systems, and environments. But as aviation and AI each evolve in unprecedented ways, it begs the question as to whether or not it’s time to reevaluate the very foundations of Human-AI teaming design and assessment.

AI-AI, Captain!

At the low end of the spectrum, AI in aviation can have fundamental functionality. On the upper end, AI-enabled aviation needs no human intervention whatsoever.

Here’s how Barry Kirwan, a safety research coordinator with the European Organization for the Safety of Air Navigation (EUROCONTROL) breaks it down:

Machine learning support already exists
Cognitive assistant equivalent to advanced automation support
Cooperative agent able to complete tasks as demanded by the operator
Collaborative agent an autonomous agent that works with human colleagues but which can take initiative and execute tasks, as well as being capable of negotiating with its human counterparts
AI executive agent where AI is running the show but with human oversight and the human can intervene
Fully autonomous AI where the human cannot intervene

Once the level of AI implemented reaches the collaborative stage it acts as an autonomous ‘teammate’ that can initiate actions, negotiate solutions, adjust priorities, and perform tasks. It seems that time has arrived. It’s somewhat similar to the old saying regarding the cost of a purchase you’re considering: ‘If you have to ask, you can’t afford it.’ Only now, if we have to ask if whether it’s time to reevaluate human-AI teaming, it’s time to do it.