AI and aviation safety: An airline perspective

Alain Corbel, Deputy Safety and Risk Management Director at HOP!, discusses how AI is shaping aviation safety, while highlighting the need for human oversight to avoid over-reliance on automated systems.

 

Artificial intelligence (AI) is transforming the aviation industry, offering new ways to enhance safety, efficiency and risk management. But as AI becomes more integrated into operations, how can airlines ensure they maintain oversight and avoid over-reliance on automated systems?

Ahead of his participation at the ERA’s Regional Airline Conference panel on digitalisation and AI, we spoke with Alain Corbel from HOP!, Air France’s regional airline, who is also Chair of ERA’s Air Safety Group. With a background in defence and extensive experience in airworthiness, flight data monitoring, and safety risk management, Corbel is at the forefront of aviation safety innovation.

Balancing innovation and risks

“The true power of AI lies in its ability to process vast amounts of data and learn from it”, says Corbel. This capability is transforming aviation safety across various sectors and as he explains, AI has the potential to positively influence all aspects of the industry, from weather prediction to maintenance and safety management.

In weather forecasting, AI is improving models to predict conditions like turbulence and icing with greater accuracy, which is becoming increasingly important as climate change affects weather patterns. Similarly, in predictive maintenance, AI analyses data from aircraft sensors, Quick Access Recorders, and maintenance systems to identify potential issues before they arise, reducing the likelihood of in-flight technical problems. Additionally, Corbel highlights safety management. AI helps make data-driven decisions by processing large sets of reporting data, identifying vulnerabilities, and detecting emerging risks more efficiently, ultimately improving overall safety protocols.

Nonetheless, with these opportunities, comes risks. “The main risk is being overconfident in AI and letting this technology become a black box without being able to understand its behaviour,” says Corbel. He notes that the key concern is losing human oversight, particularly when AI is designed to make tasks easier. “AI is here to support humans, not to replace them.”

The future of AI in aviation safety

Aviation safety relies on data, and AI is proving invaluable in improving how that data is processed. “As Chair of ERA’s Air Safety Group, what I can see for our regional airlines is how the use of AI can improve our ability to process data,” shares Corbel. He highlights the challenge of limited resources, which often means prioritising medium and high-risk areas while giving less attention to low-risk or non-safety-related data, known as the ‘green zone’.

AI, however, offers the capability to rapidly classify large volumes of reports, link them with external data sources such as weather information and ADS-B data, and provide deeper analysis. This enables safety managers to identify hazard risks and early warning signs more effectively.

“In our business, maintaining a high level of reporting is a time-consuming task as this implies providing feedback to the writers. That can be greatly facilitated by generative AI,” he explains. Generative AI can also enhance reporting by automatically prompting reporters for missing information. “This AI technology is already available on the market, and we’re entering the era of safety intelligence.”

Regulatory considerations

Europe has taken the lead in regulating AI with the EU AI Act. Additionally, the European Union Aviation Safety Agency (EASA) has published its Artificial Intelligence Roadmap (2.0, May 2023), with the goal of positioning itself as a leading oversight authority for AI while supporting European aviation industry leadership in this domain.

“In terms of our airline’s safety activities, we’re at the beginning of the AI implementation journey, and our concern is more about prioritising cybersecurity and AI training in a secure environment,” states Corbel.

As AI continues to evolve, its role in aviation safety will only grow. However, ensuring that AI remains a tool for enhancing human decision-making, rather than replacing it, will be critical to its successful and safe integration into airline operations.