How Generative AI Will Change the Commercial Aviation Industry

As artificial intelligence (AI) technologies mature and make their way into the commercial aviation industry, they could deliver a myriad of different benefits to airlines, ranging from improved flight planning to more proactive maintenance to streamlined workflows and operational efficiencies. Considering the benefits that AI could deliver to the industry, it’s no surprise that this technology was one of the significant topics of discussion at the recent Aviation Festival Asia event in Singapore.

This biennial event is one of the world’s premier aviation showcases, bringing industry leaders and decision-makers together to discuss the industry’s future and the technologies reshaping how we travel. One of the speakers at this year’s AFA was Sajin Mohamad, the Head of Solutions and APAC Customer Programs Leader at Collins Aerospace and a contributor to Connected Aviation Today.

Sajin joined a panel discussion of thought leaders at AFA to explore the future of AI in aviation. We caught up with him following the event to learn more about AI’s potential to revolutionize air travel as we know it.

Connected Aviation Today (CAT): In your opinion, how will large language models and generative AI transform the industry? 

Sajin Mohamed: Large Language Models (LLMs) are already part of several tools the commercial aviation industry uses to boost productivity and enhance customer experience. Major airlines have also deployed generative AI-based chatbots for passenger interactions.

However, the actual transformation will happen when AI capabilities are extended into other core commercial aviation processes, including revenue management, crew rostering, flight planning, and predictive maintenance.

I predict the adoption of AI for use in these core functions will increase as advancements are made in two complementary areas. The first is Responsible AI, which deals with trust, security, and risk management. The second is Compound AI, which are systems that integrate LLM with traditional methods for inference, interpretation, and retrieval to make the results more cost-effective, highly tailored, and compliant to aviation requirements.

CAT: What key AI challenges will the aviation industry face as the technology evolves?

Sajin Mohamed: The commercial aviation industry is highly regulated. This makes explainability – which is transparency into the backend AI modeling and tools – the key to safety and certification and a considerable roadblock to increased AI adoption.

“I…expect that airlines will continue to scale investments in AI tools. For example, they will invest in AI that enables smarter chatbots, agent assistance technologies, better forecasting of flight departure and arrival times, crew roster optimization, flight profile optimization, and predictive maintenance.” – Sajin Mohamed

Commercial airlines are also rightly concerned about AI tools learning from their intellectual property and enterprise data and sharing it with competitors.

While LLMs often offer the option to turn off learning, it is counterintuitive if the model cannot learn from enterprise data. To combat this, airlines may demand the ability to differentiate shareable and non-shareable data, and different hosting or prompting models that enable them to make the best use of the tools available.

While it does make sense for the industry to collaborate more on safety and operational efficiency to improve flight planning, fuel optimization, and the airworthiness of equipment across the industry, there are other areas where they don’t want to share data. Airlines will want to make sure that their commercially sensitive, competitive, customer, and financial data is protected.

CAT: How do you see LLM technology advancing in the next year or two?

Sajin Mohamed: In the next few years, I anticipate that consumer LLM will grow exponentially and we will see a dramatic increase in LLM processing capabilities. I also expect that airlines will continue to scale investments in AI tools.

“…the actual transformation will happen when AI capabilities are extended into other core commercial aviation processes, including revenue management, crew rostering, flight planning, and predictive maintenance.” – Sajin Mohamed

For example, they will invest in AI that enables smarter chatbots, agent assistance technologies, better forecasting of flight departure and arrival times, crew roster optimization, flight profile optimization, and predictive maintenance.

CAT: What about five or ten years down the road? How will AI adoption in the industry change in that time?

Sajin Mohamed: Generative AI will be integrated into more tools and offer more parameter scaling. This will generate impressive results in use cases involving computer vision, such as baggage monitoring and turnaround monitoring. I also expect multi-modal Generative AI that extends capabilities beyond just text to images, audio and video will continue to grow.  

Maturing AI will have multiple impacts on the commercial aviation industry. Airlines will be able to offer better pricing and optimize yields. They’ll be able to implement new capabilities, such as retailing and personalization, optimization of flight scheduling, slots and tail assignments, and crew optimization.

We will also see the rise of more proactive and predictive maintenance. This will result in more predictability for aircraft turnaround, and optimized resource deployment on both the landside and airside. AI pattern recognition will also enable increased biometrics adoption, which will streamline the passenger experience in the airport and make passenger flow seamless.

CAT: While some roadblocks to adoption remain, I know that AI is already making its way into the industry. What use cases already exist for AI in commercial aviation?

Sajin Mohamed: You’re completely correct. AI is not new to the aviation industry. The industry has long been using mathematical models to help forecast yield management, conduct fraud detection, and complete other complex tasks.

“AI is not new to the aviation industry. The industry has long been using mathematical models to help forecast for yield management, conduct fraud detection, and complete other complex tasks.” – Sajin Mohamed

The industry has used tools that enable predictive maintenance, predict flight arrival times, and enable flight profile optimization for many years now. These solutions have resulted in significant fuel savings, optimized usage of ground resources, and the avoidance of Aircraft On the Ground (AOG) situations and flight delays.

CAT: The EU recently finalized the AI Act. What does this Act do, and is the aviation industry ready for it?

Sajin Mohamed: The AI Act defines rules for high-risk AI use. It emphasizes the importance of transparency and explainability of AI models and human oversight of data. It also bans harmful AI practices. It was developed to encourage AI innovation while protecting fundamental rights and data privacy.

While I can’t speak for the entire aviation industry, I can say that Collins Aerospace has processes and tools in place to ensure responsible AI use. , and receive continuous consulting from our legal, compliance, and global trade teams on our engineering activities.

To learn more about the use of AI in the commercial aviation industry, click HERE.