Canada's two largest airlines say artificial intelligence can be a game-changer for aviation by helping to boost revenues, pare costs and provide passengers with a more personalized travel experience.
Air Canada and WestJet are joining airlines around the world by spending undisclosed amounts of money on AI in an effort to harness technology that promises revolutionary advantages for both carriers and passengers.
Several large airlines in the U.S. and Europe have deployed AI in chatbots that respond to common passenger questions, machine learning algorithms to help automate airline operations and facial recognition to verify identification for luggage and boarding.
"It's really an untouched area for the airline industry that we need to develop very fast," new WestJet CEO Ed Sims said in an interview, adding he'd like to use the technology to create a "virtual concierge service" similar to Amazon Alexa or Google Home.
The aviation sector's investment in AI is expected to grow from US$152 million this year to US$2.22 billion by 2025, for a compounded annual growth rate of more than 46 per cent, according to a report from research firm Markets and Markets.
New technologies like AI are essential as a doubling of global passengers over the next two decades strains air traffic control, airport and aircraft systems, according to data from the International Air Transport Association.
Many of the benefits will occur behind the scenes in airline operations, but that is expected to trickle down to passengers through improved efficiency and cost savings.
Artificial intelligence will be critical to improving the passenger experience before, during and after flights, said Rodrigo Acuna Agost, head of AI Research at Amadeus IT Group.
The global travel technology company has partnered with Air Canada to install a passenger service system that will improve service on everything from reservations to personalized customer service and simplified rebooking.
Personalizing offers based on the customer's interests and price preferences will shorten the searches travellers make before booking flights, said Steve Peterson, global travel and transportation lead for the IBM Institute for Business Value.
"Rather than give you a laundry list that you have to whittle down, it gives you a few options, all of which are very compelling because they're based on who you are and what you want," he said.
Behind the scenes, Air Canada CEO Calin Rovinescu says advanced analytics are required to keep its planes flying more than 16 hours a day to spread out its costs. AI systems can predict when maintenance is required even before a part is broken, allowing for quick repairs and minimizing time on the ground.
"That kind of data on an ongoing basis will make our entire fleet much, much more efficient and that's one of the real drivers, one of the real benefits of investing in new generation technology," he told an international aerospace conference.
Algorithms can be used to predict passenger behaviour and reduce overbooking by analyzing inputs such as historical passenger data, weather patterns and time of day to better predict how many people won't show up for a flight.
Fuel requirements can be optimized days before a flight departs by predicting baggage loads based on variables such as nationality, destination and type of traveller.
After flights, algorithms can monitor social media for passenger sentiment to understand the passenger experience.
In terms of data security, people's personal information is at less risk of being breached in aviation because most AI systems use generalized data and permissioned access to existing data such as travel history to determine which profile the passenger resembles, said Peterson.
"While the individual details about specific passengers are 'used' to draw conclusions, they do not put passenger data 'at risk' when properly configured," he said.
Aircraft manufacturers like Bombardier, Boeing and Airbus are using AI to scan reams of data to monitor their planes in flight.
So-called health monitoring on planes such as its C Series allows data to be analyzed more quickly and accurately, enabling preventative actions to be immediately conveyed to airlines, said Bombardier CEO Alain Bellemare.
AI could eventually be used on all systems of aircraft, including brakes, generators, valves, engines and avionics to extend the life of parts and minimize disruptions, offering a huge savings for operators, added Robert Dewar, vice president and general manager of the C Series program.
"Planes will be less late, they'll be more reliable and of course if they're lower cost obviously the customer benefits," he said in an interview.
AI systems using natural language processing can scan millions of maintenance logs to predict component failure and recommend fixes, says Mark Roboff of SparkCognition, which focuses most of its efforts on predictive maintenance.
"If you can shrink down maintenance-related delays to near zero for someone like Air Canada, that could save them $100 million a year, easy," he said.
The Texas-based firm, which received an investment from Boeing's Horizon X venture arm, is also testing the use of AI for augmented autopilot that would engage in urgent situations unfolding in the cockpit.