Why AI Isn’t In Airplanes – Yet
If 2024 had a slogan it might be: Powered by AI. Most tech products now emphasize artificial intelligence in their advertising. It’s quickly becoming table stakes across several industries. That’s not to say AI is everywhere, though. Some of the largest industries in the U.S. – like commercial aviation and aerospace – have taken a more measured approach to AI adoption.
The Path to AI in Aerospace
Aerospace would like to incorporate AI into their supersystems and products. The only holdup? They’re waiting for guidance from the Federal Aviation Administration (FAA), which is ultimately responsible for ensuring that commercial aviation uses AI safely.
The path to AI adoption in aerospace became clearer when the FAA published its Roadmap for AI Safety Assurance in July. The document outlines guiding principles for AI adoption such as “Focus on Safety Assurance and Safety Enhancements” and “Differentiate between Learned and Learning AI.” It also shed light on why pilots aren’t benefiting from artificial intelligence already: the industry is still aligning with regulators on voluntary consensus standards — or compliance frameworks – that certify the same implementation of AI in aviation.
The FAA writes:
The FAA is also supporting the development of acceptable voluntary consensus standards for the means to provide safety assurance of AI. This includes the application, and possible extension, of existing standards for systems (SAE ARP/4754 and RTCA/DO-254), software (RTCA DO-178C/EUROCAE ED-12), software tool qualification (RTCA DO-330/EUROCAE ED-215), and model-based design (RTCA/DO-331/EUROCAE ED-218).
The FAA is also supporting the development of new standards for AI in SAE G-34 and anticipates additional specific standards may be appropriate for different types of AI learning. There is also considerable development on safety and trust in AI outside of aviation (e.g., NIST), and FAA will participate in and leverage the use of those standards to the extent that they can be integrated into the aviation ecosystem.
The FAA suggests three roads forward for aviation companies: extend existing compliance frameworks to include the safe application of AI, develop a new AI compliance framework specific to aerospace, or adapt a non-aviation framework for AI implementation like NIST-AI-600-1 which was developed and released in July 2024 by the National Institute of Standards and Technology.
Starting with SOC 2
We’re still years away from AI copilots and LLM-generated flight paths. However, the FAA is serious about setting boundaries around AI implementation in aerospace. The agency expects to publish a draft policy statement on certifying AI in aircraft in Q42024. Aerospace companies will need to align with the FAA’s vision of AI compliance. They’ll also want to have completed foundational compliance certifications for companies that handle data (as AI inherently does).
A SOC 2 Type I or Type II certification can signal to regulators that an aerospace or aviation company is able to adhere to complex requirements and implement controls that benefit the safe use of AI in their systems. These companies will be first in line when the FAA is seeking industry input on future guidelines.
Koop’s customer assurance platform helps tech companies seamlessly navigate the complexities of business insurance, regulatory compliance, and security automation in one place.
We provide a comprehensive suite of insurance coverage that includes General Liability, Technology Errors & Omissions, Cyber Liability, and Management Liability coupled with the most cost-effective SOC 2 compliance certification on the market.
Ready to learn more? Visit our website at https://www.koop.ai or drop us a note at [email protected].