Artificial intelligence (AI) is now part of everyday conversation across business, government, education, media, and cybersecurity. Security teams increasingly use the new technology to detect threats, monitor digital assets, and respond to complex attack patterns. At the same time, threat actors use AI to scale phishing, impersonation, and domain-based attacks.
This means AI now sits on both sides of the cybersecurity equation for CISOs. On the one hand, it powers tools that can improve speed, visibility, and pattern recognition. On the other, it increases the scale and sophistication of certain cyber threats. The result is a more complex security environment where AI adoption, oversight, and risk management are developing together.
Cybersecurity leaders see AI as an opportunity
According to our survey presented in The CISO Outlook 2026 report, many cybersecurity leaders view AI as a positive force. In fact, 73% of respondents say AI is more of an opportunity than a risk for cybersecurity. A further 10% describe AI as a “strong and clear opportunity,” while 16% say it’s equally an opportunity and a risk.
It’s clear AI has moved into practical cybersecurity use cases. Security teams now use AI to support monitoring, enforcement, threat detection, and fraud prevention. According to the report, 57% of respondents use AI-based monitoring and enforcement solutions, up from 50% the previous year. Forty-four percent use AI-based solutions for threat detection and fraud prevention, up from 36%.
The increase reflects a broader need for tools that can analyze large volumes of signals across domains, networks, endpoints, and digital channels. As attack surfaces expand, AI can help identify patterns that may otherwise be difficult to detect at scale.
AI-driven cybersecurity automation requires oversight
Automation is one of the main areas where AI is contributing to cybersecurity operations. In the context of domain name system (DNS) and similar attacks, almost three-quarters (72%) of respondents, say AI-driven automation plays a protective role. However, the same finding also notes that this automation requires oversight or careful management.
While it’s true AI can support faster detection and response, leaders shouldn’t see automation as a substitute for governance or human judgment. Instead, AI-driven automation appears most useful when combined with structured controls, monitoring, and review.
Using AI-powered cyberattacks to scale
Cybercriminals commonly use AI-based techniques across multiple types of cyberattacks, including phishing, deepfakes, endpoint targeting, and domain generation.
AI-powered domain generation algorithms (DGAs) are a specific concern. Eighty-six percent of respondents say DGAs pose a threat to organizations. They can generate large numbers of domains that appear valid enough to support phishing campaigns or other malicious activity. At scale, this can make detection more difficult and increase the volume of suspicious domains that security teams need to monitor.
Impersonation also poses growing a problem, as AI helps attackers scale by increasing the volume, realism, and personalization. Respondents identified social media impersonation and defamation as the top cyber threat expected over the next three years. Employee and executive impersonation, including deepfakes, ranked as the fifth biggest area of risk. Together, these findings show that AI isn’t just changing technical attack methods but also intensifying identity-based and brand-related risks.
Third-party AI systems ignite data security concerns
One of the biggest concerns raised in our research relates to third-party AI systems. In fact, 98% of respondents are concerned about giving third-party AI-based systems and solutions access to company data. Within that group, 41% say they are “very concerned,” while 57% say they are “somewhat concerned.”
This level of concern reflects the data security questions raised by external AI tools, including large language models and other AI-based platforms. The issue isn’t limited to internal users. Suppliers and partners are also adopting AI tools, creating additional visibility and governance challenges for organizations.
Seventy-nine percent of our report’s respondents are concerned that the increase in AI tool use by suppliers and partners poses a cybersecurity risk to their organization. At the same time, risk controls across supply chains remain uneven. Seventy percent of respondents say their firms cascade risk controls to key suppliers only, while 14% extend controls to a very few suppliers, and 15% apply their organization’s risk controls to all suppliers.
AI risk management is a governance issue
The report places AI risk within a broader governance and compliance environment. AI-related regulation and guidance, including the European Union’s (EU) AI Act, the NIST AI Risk Management Framework, and the OECD AI Principles, are emerging as organizations increase their use of the new technology.
For CISOs, this adds another layer to existing cybersecurity obligations. AI governance now intersects with data protection, third-party risk management, supplier oversight, identity management, and incident response. Seventy-two percent of respondents said the level of cybersecurity threats faced by their organization in 2025 was either “critical” or “very critical,” showing that AI risk is developing within an already demanding threat environment.
AI in cybersecurity is both opportunity and risk
As a defensive tool, AI supports monitoring, enforcement, threat detection, fraud prevention, and faster pattern recognition. At the same time, AI can support phishing at scale, domain generation, deepfakes, impersonation, and more complex social engineering.
The central issue for cybersecurity leaders is balance. AI adoption is increasing, but so are concerns about oversight, third-party access, and AI-enabled attack methods. For the full findings on AI-enabled cyber risks and how CISOs are responding to them, read and download the full CISO Outlook 2026 report.
