AI in Aviation - past two weeks
Whats been happening at the AI and Aviation crossover in the past two weeks
TLDR:
EASAs AI days and what people think about ethics of Ai in Aviation
Starlux does AI ground ops
AI pilot assessment with iPhone recordings
American mulling use of AI to price tickets'
AI in aircraft design
Aviation companies in Australian AI awards
EASA holds two “AI days”
Recently, EASA has been holding an annual “AI days” event where they get a large number of people into their conference room in Cologne and talk about AI in aviation. The headliner this year was the publishing of the results of a survey on aviation professionals thoughts on the ethics of the use of AI for different applications.
I’ll go into more detail on what happened on these days in a later newsletter
Starlux does AI
Starlux - a young Taiwanese airline - is now using an AI solution from Inform to help optimise its ground operations. It includes demand forecasting, shift planning and task allocation (I assume for ground staff). This kind of thing can reduce disruptions, delays and save $$$.
AI to assess you and the instructor
The Airline Pilot Club (an AI SaaS company), is rolling out a system that can use iPhone recordings of a sim session to evaluate pilots based on the Competency Based Training and Assessment framework. They say that the AI is harsher than people because it is brutally objective.
American Airlines mulls using AI solutions to set prices
American is thinking of following Delta into using AI to set prices. The CEO talks about how its an incredibly complex business and anything that aids in decision making and increases automation will help them become more competitive.
AI Aircraft Design
Otto Aerospace - a texas based aircraft manufacturer - announces an AI tool they have developed in house to help with designing the aerodynamic confoguration of its new biz jet.
This is an area where AI can really spice things up for aviation. Traditional computational fluid dynamics calculations take huge amount of computational power, thus making them very slow, or extremely expensive. Training an AI model to do a similar job essentially gets it to capture “the vibe” of what a traditional system would do, but for a fraction of the computational cost. As far as the quality of the result - its quite easy to test and would enable rapid iterative development and potentially much better aerodynamics.
Aerospace Companies at the AI Awards
Three Australian aviation related companies have been shortlisted for the Aussie AI awards (Web Manuals (which is Swedish), Starbound Space Solutions, and Fleet Space Technologies). I will say though that Web Manuals is putting way more effort than Airbus or Boeing in sorting out the manual mess.