Brian Tossan is chief technology officer at Toronto Pearson. This opinion piece originally appeared at Forbes Technology Council.
The AI revolution is happening in split-screen. On one side, Chinese and American tech companies are making breakthroughs at an astounding pace. On the other side, businesses in the real world are moving much more slowly to deploy these new models.
Surveys show that a clear majority of CEOs believe AI will significantly impact their industry. Yet a research report from Goldman Sachs noted that despite the billions being poured into development, there is "little to show for it so far beyond reports of efficiency gains among developers."
This raises questions. Why are so many businesses still at the kicking-the-tires stage? What’s holding them back? And how do we speed up adoption? One answer, I believe, lies in the data – or rather, the inaccessibility of it.
Flight-testing AI
Aviation, the industry in which I work, provides an excellent case study for the current state of AI adoption.
Airports are a perfect environment for AI prediction. The largest hubs operate on a scale that’s difficult for most of us to truly grasp. Each sees hundreds of flights a day, transporting tens of thousands of passengers to and from destinations around the world. Every airport is a small city, and each is a node in a global network of flight routes in which disruptions can reverberate in complex ways.
Applications for AI are already being found in the skies and on the ground. Air traffic controllers in Europe and Canada use it to safely space airplanes. Some airports employ it to assign aircraft to gates and minimize taxi times. At Toronto Pearson, we are using AI to monitor aircraft turnaround and to predict maintenance needs for our 30 kilometers of baggage conveyor belts.
The common thread is that each case makes an existing process more efficient and often relies on data from only a limited number of sources. While these can be impactful – flight disruption costs airlines over $8 billion annually – as an industry, we’re not yet unlocking the truly transformative potential of AI. For that, I believe we need to change how we safely share and use data.
New connections
Consider the number of organizations you encounter at an airport: your airline, government agencies, retailers, restaurants, the airport operator, parking or transportation companies. Less noticeably, dozens – perhaps hundreds – of digital service providers are also at work supplying the systems on which the airport operates.
While each of these collects information for its own use, the extent of data sharing is currently limited. For instance, your airline knows which gate you need, while your favorite coffee chain might know you love a cappuccino in the morning. But neither alone can guide you to the right plane via the most convenient coffee outlet. The data is there, but it’s trapped in silos.
I see an opportunity for airports to step in and create information-sharing platforms. Using AI, these could provide a more seamless and personalized travel experience. For instance, by combining information on your flight and parking space availability in the airport’s lots, it could guide you to the parking zone closest to your gate on that day. If it knows your dining or shopping habits, it could take the guesswork out of finding retail and food options in enormous concourses or recommend new ones to inspire you to explore and discover. Such a platform could also work behind the scenes, potentially alerting airport hotels to flight disruptions that might cause spikes in demand.
This scenario would add a new dimension to airport functions. They could become digital marketplaces connecting passengers, airlines and services, similar to how Uber connects riders and drivers. This would create a virtuous circle: Airline traffic drives more retail business, and enhancing retail experiences can boost passenger numbers for airlines.
Data sharing between airports could enhance this further. Greater availability of real-time information would help airports optimize their operations and allocate resources efficiently. It would also spur the creation of new digital services and tools to make passengers’ travel experience smoother and more enjoyable.
Digitally, cybersecure technology is the glue that binds these elements together, safely. Creating such a system would be a complex task. Like many companies, airports operate numerous legacy IT systems that would need to be modernized. Common data languages and systems of governance would be required.
But in my view, one of the most significant challenges is not technical; it’s cultural. It requires airports and their partners to take a more expansive view of how they gather and share data. Airports would need to establish a clear vision for their data regime and embed it throughout their operations and commercial strategy.
This is where the hard work of digital transformation is needed. AI requires a data-first mindset throughout an organization. This can be a sharp departure from normal service for teams who are not used to operating in this way, so it is essential to gain trust and buy-in at the outset.
In addition, it’s important to recognize that organizations don’t share data unless there is a clear purpose. For the partner data-sharing ecosystem to work, it needs to provide mutual benefit for all parties. We need to create compelling value propositions that clearly support participants’ growth.
The technology is available. The challenge now is to make use of it.