The New Security Agenda
Summary
- AI has accelerated the discovery & exploitation of cybersecurity vulnerabilities, exemplified by OpenAI agents breaching systems & researchers finding decade-old flaws in PHP via AI assistance.
- 87% of CISOs expect cybersecurity spending to increase in 2027 driven by AI-enabled threats & an expanding attack surface from internal AI deployments.
- Security headcount is projected to rise or stay flat, with nearly half of leaders planning to increase staff to handle broader testing & faster remediation.
- AI capabilities have become the top priority for 2027, surpassing 2026 focus areas like vulnerability management & SIEM.
- Key 2027 initiatives include SOC automation, offensive security, securing internal AI agents, & AI governance.
- Only 18% of enterprises currently use AI SOC tools in production, though 45% plan new purchases for internal AI agent security within twelve months.
- Offensive security is shifting toward continuous testing, with 83% of leaders rating frequent testing as important & 44% expecting embedding in CI/CD pipelines within five years.
- The gap between finding & fixing vulnerabilities is critical, prompting demand for tools that automate remediation tickets & verify fixes via retesting.
- AI SOC agents are increasingly trusted to investigate & take low-risk actions autonomously, with comfort levels rising from 36% today to 60% in a year.
- 92% of enterprises use generative AI tools, creating risks like data leakage, permission violations, & prompt injection that are slowing adoption.
- Traditional endpoint security tools struggle to distinguish between human actions & AI agent actions, driving investment in new solutions like Neo to track agent behavior.
- AI-enhanced social engineering via deepfakes & impersonation is a top human risk, leading to the adoption of AI-powered phishing simulations & automated takedown tools.
- AI code security is a top five priority for nearly half of CISOs in 2027, up from 25% in 2026, to manage the surge in code volume.
- 44% of buyers will first seek AI security features from existing platforms, but 47% will choose startups to fill specific capability gaps.
- A noisy market with dozens of vendors per category makes evaluation difficult, causing leaders to run pilots & narrow choices to a few products for hands-on testing.
- Startups are positioned to lead the next wave of security innovation as incumbents lag in addressing emerging AI-specific security needs.
For years, the industry prepared for quantum computing to rewrite the rules of cybersecurity. Then AI got there first. This summer, OpenAI’s agents broke out of a cybersecurity evaluation and breached Hugging Face’s production systems. They exploited vulnerabilities, used stolen credentials, and worked their way into another company’s infrastructure, all to get the answers to...
- Keywords
- Growth, Infra