The industry’s challenge has always been how to make use of it. The data has existed in silos and the immensity of it overwhelmed computing power. Even where valuable insights were extracted, it took time.

But times change. Computing power has enjoyed exponential growth, and a decades-old concept has come to the fore to distil the data stockpile.

Needle in the haystack

Artificial intelligence (AI) is a much-hyped term but, says Kim Macaulay, IATA’s Chief information and Data Officer, rightly so. “AI is different from a sophisticated piece of modern software,” she says. “With software, you get out what you expect to get out—there are pre-defined terms and rules. With AI, the code adapts and the output develops. For example, ChatGPT has improved enormously since its launch in 2021 without the need for software engineers to constantly review the nuts and bolts of code and release updates.”

Making sense of the plentiful data that has always been in airline hands was much like looking for a needle in a haystack. AI is a magnet that helps airlines find that needle. Insights that might take months or even years for a human to realize can be seen by AI in a matter of minutes.

In other words, AI is essential to helping the industry extract the latent value from its data in the coming years.

AI will be seen throughout the industry, from customer-facing disciplines to back-office systems. Aviation is rules-based with standard operating procedures, making AI an extremely useful tool. Already, checkbots—kiosks to help customers in a variety of tasks—are becoming commonplace.

AI will also be critical to the success of the 100% Offer and Order environment. Personalizing a journey, especially in real time, will require AI platforms to run through the many possible permutations.

“And there’s a lot that is being done in baggage management too,” says Macaulay. “Imagine if you could accurately assess the amount of cabin baggage. It’s not just about reducing the turnaround time. Perhaps you could combine checking in a bag with lounge access, for example, creating a win-win situation for airline and customer.”

Operational optimization is another obvious target for AI. Improvements range from enhanced crew rostering to predictive maintenance. Fuel management is another major area that is receiving a lot of attention given the importance of reducing carbon emissions.

Other aviation stakeholders employing AI will also play their part. Air navigation service providers could realistically engage with aircraft on contrail and turbulence avoidance.

Finally, there is disruption management. In the complex aviation ecosystem, delays and cancellations occur for a variety of reasons and for varying lengths of time. AI can trawl through available data and come up with real time, relevant solutions at an operational or individual customer level.

What are the challenges?

There are, of course, challenges to overcome. Implementing AI—like any advanced piece of technology—is not easy and takes money and resources. Aviation is not yet at the forefront of career choices for AI specialists, something that the industry is trying to change.

Airlines must also understand where the balance between humans and AI should be. This is not about jobs or automation—AI will generally allow staff to take on more interesting roles overseeing AI output—but rather about where customers would prefer human interaction. Losing a bag or rebooking following a disruption may be areas where customers will continue to prefer human assistance. Banking, for example, closed branches when it first used AI and then realized it perhaps wasn’t the best idea. As with staff, the key is to empower customers rather than disenfranchise them.

There is a sustainability angle to consider too. Huge amounts of computing power come with a carbon cost, which will need to be properly addressed.

But, without doubt, the biggest challenge is data. “AI relies on abundant, good quality data,” says Macaulay. “But there are lots of rules about data and how it can be used, and more rules are on the way. Finding a way through this maze of regulations will not be easy.”

It’s also true that sharing data doesn’t come naturally in such a competitive industry. Governance and cybersecurity will be critical going forward. And with legacy systems still abundant, data quality is also an issue.

Getting started in AI

For airlines just embarking on the AI journey, Macaulay advises starting slow. AI is not the cure for all ills. Although this can be the perception—other “hot tech” before it, such as the Cloud and blockchain, suffered a similar fate—AI’s implementation must be carefully thought out.

Making the year of facilitation matter

Making a priority of safety, standards, and sustainability

Achieving 5% carbon reduction by 2030

“Most importantly, implementing AI must be a business-led strategy,” Macaulay says. “It is not just a technology for the IT department. AI’s ability to transform is such that its use must be discussed in the boardroom. New revenue streams, different business models, and innovative operational changes could all result from AI. In fact, airlines should think of AI as an exponential technology not an incremental one. Suddenly, new use cases will open up.”

IATA is supporting airline efforts through various channels. It has set up an AI Technical Board, comprising representatives from multiple fields, including IT, cybersecurity, legal, and business development. The aim is to allow data scientists and machine learning experts to build models in a neutral environment.

IATA is also building a data science community that is working toward diverse proofs of concept not only for airlines but also for other aviation stakeholders, such as ground service providers and travel agents. A lot of investment is also taking place in training.

“We are creating trusted research environments,” says Macaulay. “Airlines want to stay in control of their data while sharing it and we can do that now,”

Forecasting the future

To date, airlines are only seeing the tip of the iceberg when it comes to AI. Although predictive analytics seems an initial focus, particularly in the wake of the pandemic when AI might have provided crucial hints about possible strategies, AI is sure to be employed across the board.

“There is so much we don’t know yet,” Macaulay concludes. “The future is really exciting, but AI should be business-led, and we’ll see where we go from there. We must remember that data is at the heart of AI’s success and ensuring access to quality data is the first step.”

SOURCE: IATA. World Passenger Symposium program.