AI security optimism among enterprise security leaders is real and rising, but CSC’s CISO Outlook 2026 report shows it is running ahead of the controls needed to manage AI as an attack surface. CSC, an enterprise-class domain registrar and Digital Brand Services provider, surveyed 300 senior executives — including chief information security officers (CISOs), chief technology officers (CTOs), chief information officers (CIOs), and heads of cybersecurity — in Q1 2026.
- 73% of respondents say AI presents more of an opportunity than a risk for cybersecurity, yet 98% are concerned about giving third-party AI-based systems — including large language models (LLMs) — access to company data.
- 86% cite AI-powered domain generation algorithms (DGAs) as a current threat, while only 14% say they are “very confident” in their organization’s ability to mitigate domain attacks.
- 79% are concerned that suppliers’ and partners’ AI tool use poses a cybersecurity risk, yet 70% apply risk controls only to key suppliers.
- 57% already use AI-based monitoring and enforcement solutions, up from 50% last year, as AI defense adoption accelerates alongside AI-powered threat proliferation.
Where AI Security Optimism and DNS Defense Confidence Diverge
Respondents ranked domain and DNS hijacking, subdomain takeover attacks, and cybersquatting as the top three threats they faced in 2025, ahead of ransomware and malware. That ranking aligns with CSC’s core research focus on DNS infrastructure, and it sets the tension the CISO Outlook 2026 report illuminates: domain and DNS security is the attack surface respondents feel least prepared to defend, despite recognizing it as the most prevalent threat. One in ten respondents believes major businesses and organizations are “significantly underprotected” against DNS outages, and the 14% “very confident” figure for domain attack mitigation is the lowest confidence datapoint in the survey. Social media impersonation and defamation, combined with domain and DNS hijacking, are expected to pose the greatest cybersecurity threats looking beyond 2026.
The AI deployment picture is more active than passive optimism suggests. More than half (57%) of survey respondents confirmed they use AI-based monitoring and enforcement solutions, and 44% use AI-based solutions for threat detection and fraud prevention. Both figures increased year over year — from 50% and 36% respectively in 2025. The adoption trajectory is clear. But the same AI systems that security teams are deploying as defenders are creating a new supply chain exposure: 79% of CISOs are concerned that supplier and partner AI tool use poses a cybersecurity risk to their own organizations. The gap between that concern level and the 70% who apply risk controls only to key suppliers signals that third-party AI governance has not kept pace with third-party AI adoption. As Ihab Shraim, chief technology officer of CSC’s Digital Brand Services, noted in the report: “Agentic AI could further accelerate this risk by enabling bad actors to automate reconnaissance, impersonation, and domain-based attacks at scale.”
How CISOs Can Close the AI Security Governance Gap
The findings point to three concrete adjustments where CISO attention is most likely to reduce exposure, sequenced from the foundational infrastructure layer to the third-party AI risk layer.
Establish domain security as a priority alongside endpoint and cloud controls – The CISO Outlook 2026 respondents ranked DNS hijacking as a top-three threat in 2025, yet only 14% feel very confident in their organization’s domain attack defenses. Domain name system (DNS) infrastructure now qualifies as critical infrastructure, and a gap between threat ranking and defensive confidence of this size is actionable. A baseline assessment of DNS security posture — including authoritative DNS control, registry lock status, and subdomain inventory — directly targets the gap the survey surfaces. Relevant prior coverage of declining CISO confidence across multiple defense domains shows this DNS-confidence gap is not isolated to CSC’s respondent pool.
Extend third-party AI risk controls beyond key suppliers – With 79% concerned about supplier and partner AI tool use but only 70% applying risk controls to key suppliers, the gap represents a significant ungoverned population. The 98% concern level about third-party LLM access to company data is high enough to warrant extending AI-specific clauses to all supplier tiers, not just strategic ones. A tiered framework — mandatory AI tool disclosure for all suppliers, detailed controls for those with data-system access — converts the concern into a governance posture. The SANS 2026 Workforce Report identified skills-gap pressure as the top CISO constraint on implementing new governance programs, making third-party-AI policy templates a practical starting point over bespoke per-supplier assessments.
Align AI security investment with agentic AI threat modeling – The 57% AI-based monitoring adoption rate shows security teams are investing in AI defense tools, but Shraim’s note on agentic AI — bad actors automating reconnaissance, impersonation, and domain-based attacks at scale — describes a threat category that basic AI monitoring may not yet capture. Agentic AI attacks combine multiple traditional attack types (DGA, phishing, subdomain takeover) in automated sequences; detection coverage needs to span the chain, not the individual technique. CISOs who review their monitoring stack specifically for agentic AI attack patterns close the gap between the AI security optimism the survey measures and the AI supply chain confidence that the 98% third-party-AI concern figure shows is still missing.
Join our LinkedIn group Information Security Community!
