Tragedy, once again, has struck the footballing world with the discovery of Emiliano Sala’s body within the wreckage of his doomed flight, piloted by David Ibbotson. In light of recent divisions between football fans, namely the violence between Millwall and Everton, this news has undoubtedly brought fans of all clubs together in grief at losing one of their own. The circumstances regarding the crash are unclear, and due to the size of the plane (a six-seater, single piston Malibu Piper) probably never will be, as it is unlikely to have had any sort of flight data or cockpit recorder on board.

Several reasons have been posited to account for the accident, however, no solid facts have yet been obtained at the time of writing this article. Nevertheless, this unfortunate event once again thrusts the question of aviation safety into the spotlight.

So, how could we use AI to prevent air disasters and also improve customer experience more generally?

Technology could remedy many of the problems associated with aviation safety; specifically, there is considerable evidence to suggest that by increasing the automation of aircraft, and retrofitting older models with more advanced technology, tragedies such as that of Emiliano Sala’s and David Ibbotson’s could be all but eliminated.

Proponents of automation and AI have long called for automated aircraft, arguing that an automated pilot is far safer than a human pilot. Boeing estimate that 80% of all aeroplane accidents are as a result of human error. Though a plane operated by AI may lack the capacity to crack a joke over the tannoy, it would undoubtedly have: faster reaction times; a better understanding of unexpected situations; and, most importantly, would remove all traces of bravado and risk-taking from the cockpit. Indeed, by removing the arrogance and reliance on experience that many pilots have, crashes such as Alitalia Flight 404 (where the captain ignored his first officer’s correct instincts to go-around) could have been completely avoided.

Similarly, AI could be used to better interpret instrument failures. In the case of Air France Flight 447, ice on the speed sensors gave the pilots incorrect airspeed readings, leading to the plane entering an aerodynamic stall and crashing into the Atlantic Ocean, killing all 228 people on board. With the current autopilot technology installed on aircraft, the incorrect airspeed readings were not consistent with the other readings which simply caused the autopilot to disengage and hand over control to the human pilots. With advanced AI technology piloting the plane, it could have analysed the airspeed reading and understood why it was inconsistent with other readings, possibly even providing an adjusted reading based on its conclusions.

It is not just the problem of pilot error that AI resolves. AI could be used to predict maintenance issues on aeroplanes, reducing maintenance delays. 30% of delays are caused by unplanned maintenance, so, by feeding historic data of specific parts and current aeroplane diagnostic information into an AI capable machine, it could virtually eliminate unplanned maintenance by accurately predicting when parts were about to fail and when best to replace them. In fact, this is well on its way to becoming a reality. The MoD use a similar concept with their military vehicles, reducing unplanned maintenance by approximately 20%. Furthermore, approximately 500GB of data from every Boeing 787 flight is recorded and analysed to improve the design of the plane and spot any flaws. Though the extent of AI involvement is currently limited in these cases, with the development of AI it is clear for even the humblest student of technology to see that the human element in this process could be swiftly replaced by an AI machine capable of better analysing the data in a shorter amount of time and coming to more informed conclusions.

Of course, it would take an incredibly capable AI programme to facilitate this. One that, at the moment, is unlikely to appear in a cockpit due to cost considerations (it is already tricky enough making money with airlines). However, if machine learning does continue at the rate it is now, we could be seeing pilotless planes in the near future. Certainly, within most of our lifetimes.

It’s not just safety that AI could improve in the aviation industry. It could also contribute to lower prices for consumers and bigger profit margins for airline companies. Effective use of AI to analyse passenger data could easily be used to lower prices on less trafficked routes, or at undesirable times, thus ensuring planes are fuller, increasing the profit margins of aviation companies. A win-win for consumers and airline corporations alike.

It is important that the aviation industry carries on its proud history of learning from its mistakes, continually innovating, and embracing technology to provide the highest level of safety and service to its consumers. Though it may be a while before we see the complete elimination of humans from an aircraft, automation may increase to a point where humans are only required in the most unlikely of circumstances.

The use of AI and other new technologies to improve aviation safety is the future, and a (hopefully) not too distant one at that.