Network segmentation is a foundational security control imposed onto systems that default to connectivity. The field lacks an accepted metric of segmentation or efficacy study. We propose a three-part agenda: measurement, efficacy research, and the adoption of default-to-isolation architecture.
Our ability to generate congestion control algorithms (CCAs) has exceeded our ability to verify them. We argue that generation and verification should be considered together and outline three research directions for verifying CCAs generated by large language models.
This special issue examines automated programing language translation for security, evaluated against four key metrics: correctness, performance, safety, and idiomaticity. Eight articles explore agentic and large language model (LLM)-driven translation frameworks, static-analysis-based toolchains, coverage-guided testing strategies, and safe handling of tricky low-level constructs, alongside broader reflections on translation quality.
We study how changes in frontier artificial intelligence can affect the attack and defense balance, the consequences for critical infrastructure and institutional trust, and propose governance recommendations based on the moral relationships that security controls are meant to protect.
If you read the news, the artificial intelligence (AI) models are so cyber nasty and so capable of escaping control that they cannot be shown to common mortals. The reality is different: industry data shows us that we are happily sailing in a “bad AI” bubble.
