Issues shaping cybersecurity in 2027: Predictions from FortiGuard Labs
Thu, 19th Nov 2026 (Today)
Security leaders preparing for 2027 need to understand how delegated access and authority shape their exposure so they can identify where stronger controls and investment are needed. As organisations integrate AI more deeply into their operations, they grant agents access to critical resources such as email, internal documents, and business applications. Some of these agents also have authority to update records, send information, or initiate transactions. While these capabilities can improve productivity, an attacker who hijacks an agent can also exploit its legitimate permissions to steal data or manipulate business processes.
Our new FortiGuard Labs report, Cybersecurity Predictions for 2027: Emerging Threats and Evolving Defensive Priorities, examines these developments and what they mean in the year ahead. It explores how attackers will exploit increasingly capable AI systems, how criminal services will evolve, and how organisations can prepare now.
We also revisit our 2026 predictions to evaluate how events have unfolded and which trends are still emerging. By distinguishing between capabilities being demonstrated in research settings and those deployed in live criminal operations, this analysis helps security leaders understand how threats are evolving and where to best focus their defenses.
Building on the Industrialisation of cybercrime
Our 2026 predictions report examined the industrialisation of cybercrime and the accelerating attack lifecycle. We anticipated that greater automation and specialisation would help criminals expand their capabilities and operate more efficiently, and developments observed in 2026 support that assessment. These themes are even more relevant today as AI becomes increasingly embedded in business operations and criminal workflows.
Our 2027 outlook builds on that foundation, comparing last year's predictions with observed threat activity and exploring how emerging developments could shift attackers' targets, the tools and services they acquire, and how organisations must begin preparing now. This includes examining how criminals will combine stolen access, specialised AI services, and more adaptive attack methods to expand their operations. For defenders, these developments call for closer scrutiny of the authority granted to automated systems, along with practical testing of their ability to detect misuse, disrupt compromised workflows, and revoke access.
A growing contest over trusted access
When organisations authorise AI agents to retrieve information, use applications, and perform tasks, they also grant them access that attackers may seek to exploit. As a result, an agent can become a valuable target even without a conventional software vulnerability. The consequences depend on what the agent can access, what actions it can take, and whether it remains aligned with its assigned task.
These relationships extend security responsibilities beyond identity verification. Organisations need to understand how authority flows between users, agents, and connected services, and how to detect and interrupt compromised workflows. The information, integrations, and persistent context AI systems rely on add further considerations for securing access.
The economics behind emerging AI threats
Criminal adoption of AI will depend in part on whether tools and services produce useful results at an attractive cost. Access to AI resources has value, as does the ability to package models, tools, and supporting services into capabilities that other operators can purchase. Greater specialisation could help participants obtain resources, adapt capabilities, and support their "customers" without building every component from scratch.
These commercial incentives could shape the next phase of AI-enabled cybercrime, but advertised capabilities don't always translate into reliable services. Defenders need to distinguish between offerings that could significantly expand criminal capacity and those that still face substantial practical barriers. Our predictions examine how this market is likely to evolve and what would make its services useful to criminal operators.
Speed and adaptability require closer examination
AI-assisted analysis and automation will further reduce the time defenders have to assess vulnerability disclosures and protect exposed systems. Attackers will be able to coordinate more of the process of identifying targets, adapting available tools, and testing an approach. Emerging research also points to systems that can evaluate results and adjust their actions after an unsuccessful attempt.
Progress on individual tasks does not yet establish that a fully autonomous attack can operate reliably in unfamiliar environments. Practical feasibility depends on model capability, computing resources, access, and operational constraints. Our predictions consider these factors together, distinguishing developments organisations should prepare for in 2027 from possibilities that may take longer to mature.
Connecting the outlook to defensive action
In 2027, organisations must retain control over delegated authority while improving their ability to turn threat intelligence into timely protection. This requires understanding their actual exposure, assigning clear responsibility for decisions, and testing the effectiveness of protective actions. As security teams automate more analysis and response, they also need confidence that the information their systems examine won't misdirect their actions.
These priorities also extend to the broader security community. Criminal operations rely on infrastructure, service providers, financial relationships, and recruitment. Sustained cooperation can help defenders address these dependencies and make it harder for operators to recover from disruption. Prevention and early intervention can also reduce recruitment, helping limit criminal networks' ability to replenish their capabilities.
The 2027 cybersecurity predictions report brings these considerations together to guide security planning and investment. It evaluates emerging threats, shares the evidence supporting our forecasts, and identifies the defensive actions organisations should prioritise.
Read the Cybersecurity Predictions for 2027: Emerging Threats and Evolving Defensive Priorities report from FortiGuard Labs to explore the full predictions and recommended actions for the year ahead.