AI Slowdown Unlikely to Impact Cybersecurity, Experts Say

There's been talk about slowing down AI development due to concerns about its impact on society. However, experts at Cisco Talos argue that an 'AI slowdown' is unlikely to affect cybersecurity.

The main reason for this is that AI models have already reached a point of incremental improvements in cybersecurity capabilities. In fact, the newest models are getting better so fast that our ability to use them effectively for defensive tasks hasn't kept up.

On the offensive side, most recent AI models can mine decades of tech debt to uncover vulnerabilities. Instead of chasing model improvements, experts suggest improving agentic harnesses and frameworks to get better capabilities with existing models.

The focus should be on addressing other security problems such as asset and role inventories, identity management, least privilege, and segmented networks. These fundamentals are more impactful in making it harder for threats both human and agentic to successfully carry out attacks.