AI has moved from experimentation to everyday use in embedded systems. Teams are now relying on AI-generated code in products that run medical devices, industrial equipment, vehicles, and energy systems. But the security practices surrounding that code need to evolve to keep pace.
This report shares insights from more than 200 embedded systems professionals working across critical infrastructure sectors. It examines how AI is being used in embedded development today, the risks teams are seeing, and the security gaps that remain as AI-written code moves into production.
The 2025 AI in Embedded Systems Report highlights the challenges, trade-offs, and priorities shaping the next phase of embedded security and where organizations need to focus to keep critical systems safe.
Download the report to access all the findings and recommendations.
Key Takeaways: The FDA is asking medical device manufacturers to submit VEX (Vulnerability Exploitability eXchange) files alongside SBOMs in some premarket cybersecurity submissions. VEX artifacts document whether known vulnerabilities in SBOM-listed components are...
URGENT/11 vulnerabilities allow attackers to take control of embedded devices remotely without any user interaction and often without triggering traditional security defenses. Discovered in 2019, these eleven flaws in the IPnet TCP/IP stack affect millions of devices...
Most of the code running in your product probably isn't code you wrote. Open source libraries, vendor SDKs, and firmware components all become part of your attack surface the moment they're compiled into your build. Third-party software risk is the potential for...