Making hardware trojans using AI: artificial intelligence in hardware design and cybersecurity. 📟🤖💉☣️🎰
An international group of academic hardware security experts shared their work last year on how new types of artificial intelligence, especially large language models (LLMs), may help design new microchips and hardware components.
An especially interesting part (for me) is how AI can contribute to hardware security: models and AI-enabled systems can not only design new architectures but also identify weak points and the presence of hardware trojans.
Actually, AI can also identify ideal places and propose the design of hidden malicious circuits - hardware trojans. The paper discuss “test trojans” for debugging purposes, but it’s a matter of how they are used.
Super interesting and very insightful work, full of surprises and ideas for future research. Enjoy the read, and don’t forget to share. Thanks!
More details:
Hardware Design and Security Needs Attention: From Survey to Path Forward [PDF]: https://arxiv.org/abs/2504.08854
目录
最新
- Hacking self-driving cars in the real world: laser pointer attacks & drones that make cars disappear. 🚗⚡😎🚨💥
- Reverse engineering of the Fermax intercom system: ongoing hardware security research. 🚪📸🔬 👨🏻💻🧰
- Hacking with malicious USB drives or malicious QR codes? The most effective way to break in. 🥷🏻💾𝄃𝄃𝄂𝄂𝄀𝄁𝄃𝄂𝄂𝄃🧱🏆
- How to become a Android 𝗁̶𝖺̶𝖼̶𝗄̶𝖾̶𝗋̶ security researcher: malware, exploitation and more. 🎓🔬👨🏻🏭🧠👨🏻💻
- Hacking a Linux device with TPM + LUKS encryption: extracting the TPM key from SPI in plaintext. 🐧💻👂🔑🎉
- Hacking a GE fridge with a hardened STM32: Bypassing new protections to extract flash memory. 🧊🪛👨💻🪠💾
- Security & privacy in the Tile tracking system: the first (public) security analysis of its protocol. 🐾🗺️🎯👨🏻💻🛠️
- Hacking using AI-written exploits: human tests of LLM agents’ PoCs show they are almost all fake. 🤖🖋️📜👨🏻💻🤡