AI Slowdown or Not, Cybersecurity Basics Still Matter Most

There's growing discussion about deliberately slowing the pace of AI development due to ethical, safety and geopolitical concerns. According to Cisco Talos, whatever the merits of that debate, an AI slowdown would likely have little effect on cybersecurity specifically. AI models are already highly capable, meaning newer versions tend to bring only incremental gains for defenders, while attackers can already use existing models to uncover large numbers of vulnerabilities in ageing systems.

Talos suggests that rather than waiting for better models, organisations may get more value from improving how they use the AI tools they already have, through better frameworks and processes. More importantly, many businesses still haven't mastered basic cybersecurity fundamentals such as keeping accurate asset and role inventories, managing identities properly, enforcing least privilege access, and segmenting networks.

These fundamentals may not be as exciting as AI, but they do more to stop both human attackers and AI-driven threats from succeeding. Talos is clear that businesses shouldn't ignore AI while fixing these basics, but they shouldn't rely on AI alone either.