AI impacts in aviation – Wings MagazineWings Magazine

The past century has been, in large part, fuelled by the global growth of aviation and aerospace, which only accelerated out of the 1980s proliferation in computing with the innovation behind microprocessors, RAM, storage and networks. Today, artificial intelligence (AI) presents a new business paradigm in its ability to review and analyze millions of data points and provide context. For the non-technical, AI is the simulation of human intelligence by machines and computer systems, allowing them to learn, reason, solve problems and perceive. AI systems can process and analyze copious amounts of data to recognize patterns, make predictions and decisions. While AI is commonly used as a digital assistant, it can also be used to enhance an organization’s operations. AI can create code, develop procedural steps, serve as an agent, optimize workflows and provide metadata and advanced analysis. 

The Canadian federal government’s 2024 budget included: investing $2 billion to build and provide compute and technological infrastructure for AI researchers, startups and scale-ups; enhancing AI startups through $200 million in support from Canada’s regional development agencies; investing $100 million into the National Research Council Canada’s (NRC) AI Assist Program; supporting workers who may be impacted by AI with $50 million over four years through the Sectoral Workforce Solutions Program; creating a new Canadian AI Safety Institute with a $50 million investment; and providing $3.5 million over two years to advance Canada’s leadership in the Global Partnership on AI.

The FAA has been thinking and planning for AI in aviation since the early 2020s. Dr. Trung Pham, Chief Scientist and Technical Advisor, FAA, reports: “Artificial Intelligence is the discipline of creating computational systems that mimic aspects of human intelligent capability to perceive, decide, and act. In aviation, AI systems are implemented to enhance the effectiveness and efficiency of controlling aircraft systems. Machine Learning (ML), often considered a key subset of AI, applies computational methods to train AI models to learn from data and generalize that knowledge into compact algorithms for implementation in code. While theoretical research in ML advances computational technique, aviation research focuses on measuring the functionality and performance of AI systems in accordance with the certification framework to assure the safety of aircrafts containing AI systems. Discipline leadership supports evaluating the effective use of ML and safely integrating AI technologies in aviation systems, informing FAA policy, guidance, and training. This effort includes international collaboration with industry, other government agencies, standards development organizations, and academic institutions to advance understanding of algorithm development, data characteristics, and model functionality and performance.”

There are a number of AI enabled systems being developed by the FAA, including (but not limited to): Strategic Management of Airspace Routing Trajectories; AI Enabled ATC Decision Support; Drone Identification Tool; Safety Data Analysis; Maintenance, monitoring and control of the National Airspace System; AI safety assurance methods, which allow for the safe introduction of algorithms into aviation; and other applications.

Bombardier Inc. has been using AI to increase productivity and enable condition based and predictive maintenance, while reducing aircraft downtime. Bombardier’s AI driven tool, Smart Link Plus, transforms an aircraft’s data into actionable, proactive and prescriptive solutions, allowing operators to leverage the power of their aircraft’s data to minimize downtime. 

Internally, Bombardier employees use AI tools to quickly find and understand information across a complex aerospace knowledge base (spanning engineering standards, maintenance manuals, quality procedures and evolving operational documentation). This is especially valuable in aviation, where details are technical, highly regulated, version controlled, and distributed across many systems; and where accuracy and context matter as much as speed. Bombardier is applying AI internally through enterprise software tools that help improve productivity and support better-informed decisions. One example is CLARA, an internal sales tool that helps streamline the creation of range maps and supports more efficient customer discussions. On the engineering side, Bombardier recently signed a multi-year agreement with CoLab AI Inc. to deploy artificial intelligence solutions that will support the design and manufacturing processes of its business jets. With this collaboration, Bombardier will harness AI to drive innovation throughout product development cycles, enabling faster timelines and adding advanced AI driven capabilities to current procedures. Bombardier continues to explore the use of AI to improve manufacturing processes by reducing aircraft cycle time, enhancing product quality, and identifying production bottlenecks earlier, while driving greater efficiency across operations.

Savvy Aviation has many products which utilize AI. Its first AI project focused on failing exhaust valve analytics (called FEVA). FEVA was created in 2014 and utilizes AI to report on signs of a burned exhaust valve in the company’s database. On average, Savvy found one cylinder in 40 would have a burned exhaust valve, but that cylinders, reported by FEVA “at elevated risk”, have roughly one chance in four of having a burned valve. Savvy recommends that any cylinders reported to be at elevated risk by FEVA be borescoped at the next opportunity. Its machine learning language was trained using hundreds of thousands of flights. More recently, Savvy’s GADfly technology utilizes an AI agent to review data from Cirrus SR22 aircraft that has been uploaded to Savvy. The difficulty in Savvy’s nine million flights (in its database) is unlabelled data, which can be challenging for AI engines. 

Dr. John Sipple joined Savvy’s team to resolve this problem of unlabeled data, creating a technique called MADI (Multivariate Anomaly Detection with Interpretability). GADfly provides blame analysis that identifies which specific data element (of a flight) contributes the most to an anomaly score for a Cirrus aircraft, which Savvy staff can use to escalate (to the aircraft owner) future maintenance issues before they occur. Mike Busch, founder and CEO of Savvy, reports: “We are also working on AI projects in other areas of Savvy. The Savvy platform has a huge database of all sorts of things, like 250,000 borescope images, thousands of maintenance records with many handwritten records and stickers. Our newest project, Ask Anything, utilizes typed commands to search through all the records of our database, delivering an AI super search capability to find the relevant records in the database. I am thrilled to be on the ground floor of AI, which will massively change everything people do. These powerful AI tools are getting demonstrably better every week.”

North America is in the early stages of utilizing AI on the operations side of an airport. Asia Pacific is further ahead, implementing AI-based systems to improve efficiency and safety. Alan Tay of Ilumia Labs reports: “My company has been working the embedding AI systems in aviation for the past two years,” says Alan Tay of Ilumia Labs. “In the Asia Pacific, we see airports implementing AI into Gate Management systems, realizing a five- to seven-minute [per hour] efficiency gain, allowing the gates to support more aircraft per hour. Some major airlines in the Asia Pacific have embedded AI and facial recognition into their gate boarding systems, eliminating the need to scan boarding passes. The result is 248 to 336 passengers boarding a Boeing 787 Dreamliner in 30 minutes. This has the long-term effect of allowing the total aircraft accommodated per hour to increase, thereby allowing the airport to avoid or defer airport construction projects. Subsequently, annual movements are now limited by weather events and aircrew shortages.” 

AI augmented systems have been implemented across the aviation ecosystem. One example is AXIS Flight Simulation unveiling its AI Debriefing Station in 2025, an advanced automated platform designed to improve pilot training, support instructors and standardize assessments. CAE is focused on AI development, with the release of CAE Rise. In 2022, CAE announced that Air Busan had implemented CAE Rise. Nick Leontidis, CAE Group President, Civil Aviation, reported: “Air Busan and CAE share a longstanding partnership that spans almost a decade and we are thrilled to expand the CAE Rise training system to their Airbus A320 fleet. With CAE Rise, Air Busan can now harness the power of big data analytics to train the safest pilots, while leveraging insights to continuously adapt and improve Air Busan’s flight training programs.” 

Iridium Communications Inc. and Aireon LLC are both using AI after making a substantial investment into AI implementation. Aireon created a walled garden with its first AI enablement focusing on cyber security monitoring and implementation in its 24/7 customer service environment. With GPS jamming and spoofing on the rise, they see the utilization of AI based analytics into their robust databases, which have been collecting data since 2017. There are approximately 16,000 aircraft in the air at any given time transmitting ADS-B and acting as sensors for their space-based ADS-B system, which, with AI, can enhance GPS spoofing/jamming products and turbulence avoidance technology. 

Bombardier in April 2026 signed a multiyear, multimillion-dollar contract to use artificial-intelligence software from CoLab AI to help with aircraft design and engineering.
PHOTO: Bombardier

“We have made substantial AI investments in our team and infrastructure and have created a secure enclave for our global ADS-B data and now are exploring AI-based applications, including looking at how to apply it to turbulence detection and prediction applications,” says Don Thoma, Chief Executive Officer, Aireon. “We are also working with partners who are building AI-based air traffic analysis and airspace management tools, which is a key part of our early strategy to use the power of AI tools and very unique datasets to improve safety and efficiency for aviation stakeholders.” 

Matt Desch, Chief Executive Officer of Iridium Communications, explains: “AI is being integrated into Aireon products and we expect them to be able to mine important real-time and historical insights from their very rich and high-quality dataset. Their commercial data services will be turbocharged by AI, creating revenue and increased profitability. We don’t need AI to generate cost reductions, but it’s more about creating new products and services. Success with AI will be about the quality of data and Aireon is the gold standard for global aircraft surveillance and movement information.” 

Companies like Jeppesen ForeFlight have been using AI in its applications for years. Henrik Hansen, Chief Technology Officer, Foreflight, reports: “Specific to aviation, I’m confident it will change our industry for the better, it’s just a matter of when. When considering the pace of AI adoption, I think about it in terms of three factors: optimization, regulation and trust.” ForeFlight’s first step in leveraging AI in its app is enhancing large language models (LLM) to save users time in their flight planning. ForeFlight has trained LLMs to analyze user airport comments and summarize them for quick reference. This process is executed in ForeFlight’s cloud ensuring that ForeFlight’s consistency and traceability on the iPhone or iPad is not compromised. 

In 2025, Jeppesen ForeFlight added Ryan Walker as Head of AI and Portfolio Marketing to the team. “I see the biggest opportunity of AI in bridging the gap between ATC, operations and the cockpit through continuously updating both ends of a flight,” says Walker. “Disconnected operations are costing the airline industry over US$60 billion annually. However, there is risk in over confidence in systems not grounded in certified data. With certified systems and data, AI analysis and problem solving should not deliver wrong answers.” 

Travis Root, Senior Vice President, Flight Deck, Jeppesen ForeFlight, reports: “Safety is the biggest opportunity for AI to connect teams, which for business aviation and the airlines is a natural occurrence. General aviation pilots tend to be a team of one, however, AI delivers a team of experts for every pilot. As the pilot is the final decision authority, there is risk that in the early days of adoption, AI can be wrong, requiring guard rails to manage expectations and help the pilot make the right decision.”

NAV CANADA is taking a pragmatic and gradual approach to AI, focused on three main areas. The organization is progressively operationalizing the technology, starting with business functions like human resources and finance, and extending into operational decision-support tools that complement, rather than replace, the work of its controllers and technical staff. “We are in the early days of AI, with NAV CANADA operating in a heavily regulated and safety-focused industry. As such, we are taking a controlled deployment strategy designed to ensure that safety is not compromised,” says David Sheppard, VP and Chief Technology and Information Officer. “However, we believe the benefits to NAV CANADA will be profound in the long run, delivering productivity enhancements across our workforce and workloads. Where AI is a component, our analysis already illustrates productivity gains.” 

As RPAS operations and Beyond Visual Line of Sight (BVLOS) commercial flights grow in Canada, AI is one of the enabling technologies that will help scale simultaneous operations without overloading air traffic management systems (ATM). Today’s ATM is built around human-to-human communications, with technology supporting controllers and pilots on either end. The RPAS segment is structurally different, relying on richer data exchanges, evolving standards, and lower-cost hardware, which allows for faster integration of new technology than the longer certification cycles of conventional aviation typically permit.

Both the U.S. and Canada, as well as other countries, are evaluating AI-based unmanned aircraft designed to pair with manned aircraft, whereby the unmanned fighter aircraft will (in phase 1) be autonomously controlled from the ground, but in phase two will be controlled by AI systems. With the ability to currently target up to 100 enemy points, the only limitation is the number of weapons on each autonomous AI-driven attack aircraft. With the issues that NATO-based countries have been experiencing over the past year(s), however, there is an obligation of each country to ensure that the AI systems (delivered by foreign companies) have no hidden code, which allows that country to lose control of the attack vehicle to a foreign country. An obvious question is will AI systems or human IT experts be used to audit the AI attack systems for compliance to country specific requirements.

In 2024, ICAO released a series of seminars on ICAO TV discussing the utilization (both present and future) of AI in aviation. There are numerous episodes with many hours of content, which provides information about AI from a variety of stakeholders. These episodes are a great way to gather information from organizations on the leading edge of AI implementation, providing a real-world assessment of the benefits and drawbacks of the path to AI enablement.

Aviation-based AI systems can change the results of every flight through ensuring that the human aspect of the flight does not compromise passengers and aircrew through bad decisions. One might argue that (over time), mid-air and ground collisions will become a thing of the past, as more AI integration into avionics, air traffic management (and control) and airport automation systems take place, preventing (or reducing) bad decisions. In-cockpit AI-driven technologies may hold the potential to reduce the number of accidents and incidents. Which would be priceless. | W