For years, attackers have held the advantage.

They only need to find one weakness, while security teams are responsible for protecting everything. But according to Gartner's 2027 cybersecurity outlook, AI could begin changing that equation.

At this week's Gartner IT Symposium, Distinguished VP Analyst Leigh McMullen made a surprisingly optimistic case for how automation could shift the advantage toward defenders. But not everyone is convinced the shift will happen overnight.

One CIO we spoke with believes AI will narrow the gap, but attackers will maintain their advantage in the near term. After all, attackers only need to be right once, and they're adopting AI faster and with less friction than enterprises can govern it.

The conversation also went beyond who has the upper hand. Gartner challenged how organizations measure cybersecurity success, placing greater emphasis on proven recovery capabilities rather than prevention alone.

As one CIO pointed out, boards want to know how quickly operations can be restored, not whether every attack was prevented.

Read the full story for Gartner's 2027 outlook and perspectives from CIOs on what it will take to turn AI into a defensive advantage.

According to Semperis’ 2026 Ransomware Risk Report, AI was involved in 84% of reported ransomware attacks, including 44% in which attackers exploited organizations’ own AI agents or systems. Identity systems were targeted in 87% of incidents, yet fewer than half of affected organizations had automated recovery tools. Nearly 7 in 10 organizations also reported experiencing multiple successful ransomware attacks.

Meanwhile, Proofpoint’s 2026 survey found that 78% of CISOs expect to manage AI risks without additional resources, while 77% report excessive expectations. The findings come as the CISO’s role increasingly covers business resilience and AI governance, with authority and resources struggling to keep pace.

Hertz has appointed John Koester as Senior Vice President and Chief Information Security Officer. Koester joins the company after serving as Vice President and CISO at Graphic Packaging International for two years. Keep reading…

By 2030, Gartner predicts 80% of organizations operating public-facing AI will experience cost-exhaustion attacks designed to drive up consumption and expenses. The firm also expects more than 10 million autonomous AI agents to clog public services by the end of 2027, recommending guardian agents that can identify and stop unauthorized activity.

Meanwhile, Gartner's top strategic technology trends for 2027 predict that 25% of enterprise security breaches will trace back to AI agent abuse by 2028. AI agent identity and autonomous exposure remediation were among the technologies highlighted, with an emphasis on managing agent privileges and using automation to identify and resolve vulnerabilities.

Two critical vulnerabilities in PHP have prompted emergency security updates and temporary website shutdowns worldwide.

Known as PHP2Shell, the flaws could allow attackers to run malicious code on vulnerable web servers without logging in. Censys identified nearly 55 million web properties running PHP when the vulnerabilities were disclosed, although the number exposed to these attacks remains unknown.

The weaknesses had existed for decades but were difficult to exploit using traditional methods. AI-assisted research involving OpenAI and cybersecurity partners helped uncover techniques that made exploitation possible.

Emergency patches were released on October 21, but some organizations had already taken services offline. Cloudflare reported millions of exploitation attempts, while Amazon identified suspicious activity targeting PHP2Shell four days before the public announcement.