How to Steal AI Algorithms at the Edge and Protect Them

Artificial Intelligence (AI) algorithms are increasingly deployed in edge environments, raising intellectual property (IP) protection concerns due to heightened risks of tampering, reverse engineering, or theft of sensitive models and data. While board-level hacking is common and well-known, microelectronics tampering is rarer, more complex, and harder to detect — its success can severely compromise systems and diminish competitive advantage. Learn what applications are vulnerable to microelectronics attacks, methods and risks of microelectronics tampering such as side channel and beam attacks, why commercial security engineering limitations can lead to national security threats, and how Mercury is pioneering secure microelectronics for AI and IP protection.

Discover:

• How your IP, AI tech, data, and sensors aren’t as safe as you think.
• Commercial and defense applications that can be at risk: Tesla and MQ9
• Why trust and assurance are more than just domestic manufacturing
• Microelectronics tampering methods: imaging, SCA, beam attacks, delayering, probing, and more
• Risks of microelectronics tampering methods
• Limitations of commercial security engineering and how corporate espionage can become a national security threat
• What secure microelectronics are and how Mercury is ensuring their resiliency