Particle.news
Download on the App Store

AI Security Emerges as the Bottleneck, Cisco Summit and New BSIMM Report Find

Leaders at Cisco's San Francisco summit said trust-focused security now gates the shift from pilots to production.

Overview

  • Executives from Cisco, OpenAI and AWS said enterprises are holding back deployments due to a trust deficit, with Sam Altman calling for a new security and data-access paradigm.
  • Black Duck’s BSIMM survey reported that securing AI-generated code is now the top application-security priority, with a 12% rise in teams risk‑ranking LLM code and a 10% increase in custom rules to catch AI-specific flaws.
  • Cisco’s Jeetu Patel identified three constraints slowing adoption: limited infrastructure capacity, a trust deficit and an emerging data gap as public human-created training data is exhausted.
  • Cisco signaled major infrastructure outlays, with Patel saying the company will spend billions to expand compute, power, memory and network bandwidth for AI workloads.
  • Security governance is shaping buying decisions, including rising sovereignty requirements, while AWS’s Matt Garman urged clearer success criteria and change management to move projects from proofs of concept to production.