Dive Brief:
- As the growth of AI increases pressure on the infrastructure of enterprise networks, IT executives are deeply concerned about the cybersecurity consequences of that frenzy of network change and expansion.
- “Respondents cited security complexity as the most significant challenge associated with AI-driven network demand,” Cisco said in a recent report about that network demand.
- Security concerns represent “a direct barrier to AI scale,” the report found, because organizations don’t want to expand their use of AI if they believe doing so creates hacking risks.
Dive Insight:
Cisco’s report offers a window into the AI security concerns shaping enterprise adoption, as well as highlights organizations’ expectations about the future of AI-powered cyberattacks.
Cybersecurity concerns were one of the biggest reasons why organizations felt they needed to modernize their networks to support AI tools, with 72% of IT leaders citing “increased security risks or an expanded attack surface” as a motivating factor.
At the same time, more than three-quarters of respondents to Cisco’s survey said they expect the security risks of AI to increase as companies expand their use of the technology beyond generative activities. A similar share of respondents said AI “has already expanded their attack surface in the last 12 months.”
Another worrisome finding: 71% of IT leaders expect AI threat evolution to outpace existing security controls.
“We’re just playing catch-up at the moment,” a senior IT executive in the U.K. education industry told the interviewers.
Nearly 70% of respondents told Cisco that they were seeing an increasing number of blind spots on their networks, limiting their ability to monitor and block suspicious activity.
Cisco’s report is based on a survey of 3,472 CIOs and other technology leaders around the world between March and April 2026.
“Respondents pointed to expanded attack surfaces, shadow AI activity, inconsistent policy enforcement, and limited visibility into how AI-driven traffic moves across enterprise environments as key factors driving [AI] hesitation,” Cisco analysts wrote.
In some organizations, the sprawl of AI tools has stressed security teams to their limits. One retail-sector executive told interviewers, “The issue from a security standpoint is that it’s hard to create the guardrails for every possible AI tool that your organization must use.”
Nearly 90% of IT leaders report adding security controls to their AI tools, but many of them aren’t sure those controls will be enough — 61% said they were waiting for “greater confidence in their security posture” before expanding their organizations’ use of AI.
Organizations using AI face a myriad of balancing acts involving security, speed, efficiency, and control, as AI tools operate at machine speed, agents access sensitive data and shadow AI lurks outside security teams’ reach.
“Gaps that once caused minor operational issues can now create significant governance and security exposure inside environments where AI systems continuously generate activity across distributed networks,” Cisco said.
