Can AI-Assisted Avionics Bolster Business Aviation Safety?

April 29, 2025

With an aim toward maintaining safety, artificial intelligence has great potential to significantly assist pilots, say experts, although humans are expected to remain the ultimate decision-makers on the flight deck for the foreseeable future.

Specifically, recent AI advancements have bolstered flight deck safety by accelerating efficiency, reducing pilot workload and increasing operational preparedness. Several OEMs are developing AI technology in the avionics space, including Airbus, Boeing, Collins Aerospace and Honeywell. Honeywell’s core avionics system Anthem, unveiled nearly four years ago at the 2021 NBAA Business Aviation Convention & Exhibition, is designed to provide a seamless integration of so-called “deterministic AI” into the flight deck.

Anticipation is one of Anthem’s key features, said Jeff Woirhaye, senior director for Honeywell Anthem at Honeywell Aerospace Technologies.

“The pilot begins to type into the scratchpad, and the system anticipates where the information needs to go based on context and brings up the correct field,” Woirhaye said. “This innovation allows the Anthem cockpit to stay in sync with the pilot and improve their efficiency inflight.”

Honeywell defines deterministic AI as systems that always produce the same output for a given input under the same conditions. Under this notion, its AI also streamlines the interaction of maintainers with their ground applications, where a simple input gets converted to adaptive triggers to accomplish multiple preflight tasks.

“Similarly, for en route and post-flight phases, AI and its explainability are key to enhance directed troubleshooting and anomaly detection using streamed/filed maintenance data,” Woirhaye said.

AI and Pilot Decision-Making

Along similar lines, while AI-powered avionics aim to improve safety and efficiency, aviation is full of unexpected variables.

For Kevin Singh, instructor pilot for the Global Vision program and founder and president of Icarus Jet, balancing trust in automation with real-time human decision-making is paramount.

“The balance is tilted towards a safer sky by having a well-trained ATC, pilot, engineer, cabin crews and automation, which help with decision making, but don’t actually make the decisions,” Singh said.

Experts also point out that the potential for AI-driven avionics to enhance a pilot’s ability to manage complex scenarios must be balanced against possible pitfalls brought by overreliance on automation.

Honeywell has measured the impact of Anthem’s AI functionalities through substantial Voice of the Customer exercises, pilot demonstrations and evaluations.
Feedback received during these sessions has been very positive, said Woirhaye. This strategy is also being applied to Anthem, addressing the FMS Performance Prediction function challenges from heavy oversight with human reviewers, human factors and flight tests.

“At Honeywell, we put a huge emphasis on building that trust and clarity,” Woirhaye said. “The human is always the captain – the ultimate decision-maker – and the AI’s job is to support and augment them.”

Evolving AI-Assisted Avionics

Anthem is designed to support growing levels of aircraft autonomy. That means it can handle more tasks than pilots traditionally do – if and when regulations and certifications allow, according to Honeywell. Today, that might be automated engine startup checklists or route optimization. Tomorrow, it could manage entire flight phases with minimal human input.

“We’ve deliberately made the system scalable and upgradable, so as computing power increases or new algorithms are developed,” Anthem can take advantage, said Woirhaye.

Additionally, pilots can set their own preferences for configuring the displays, which gives a more natural feel during critical phases of flight.

“Anthem’s PilotPredict feature anticipates the pilot’s actions,” Woirhaye said. “Items like airspeed, heading and radio frequency, among others, are naturally populated on the flight deck.”

How Business Aviation May Capitalize on AI

Singh said business aviation will likely benefit most from AI. He sees big opportunities for AI to assist pilots in many areas, including those that have safety aspects, such as:

• Aircraft performance calculations
• Flight planning and rerouting
• Weather deviations and avoidance of clear air turbulence
• Fatigue and sleep management
• Aircraft maintenance scheduling
• Identifying false crew alerts

Maintaining Cybersecurity in Aviation AI

However, with increasing inflight connectivity and data reliance, cybersecurity challenges might arise and potentially hinder the integrity of AI-driven systems.
Woirhaye said Honeywell has a very strict posture on cybersecurity. Its onboard connectivity gateway is designed to be compliant with DO-326B (the Radio Technical Commission for Aeronautics [RTCA] document that provides an Airworthiness Security Process Specification for aviation).

“This secure architecture helps us ensure the integrity of the systems that are interacting with the avionics,” Woirhaye said. “While the data integrity risks, their impact and mitigation have always been part of avionics development, it doesn’t matter if it’s an AI-driven system or non-AI-drive system.”

A Crucial Approach to Pilot/AI Interfacing

In a nutshell, if Honeywell’s AI suggests a course of action, the company wants the pilot to understand why and then ensure that it makes sense. Over time, as AI proves itself reliable and helpful, trust grows naturally, said Woirhaye. “This approach will be crucial as more autonomous capabilities come into play.”

“We anticipate technologies like advanced speech recognition, computer vision and even machine learning-based weather prediction will play a role on the flight deck of the future,” Woirhaye said.

Singh expects that while AI will continue gaining ground in the business aviation industry, the elements of traditional flight training and experience will remain vital on an AI-assisted flight deck.

“Flying will inherently not change due to AI,” Singh said. “Thrust will still be produced and lift acquired with the motion of the aerodynamics. Primary and secondary controls will still control the movement; and training will be done via virtual reality; and a deeper visualization of parts and functions will be provided to the next generation of pilots.”