Security organisations are increasingly confronting threats that develop faster than the evidence available to assess them. As artificial intelligence gains greater autonomy, established assumptions about cyber operations, oversight and organizational security are being tested alongside the need for better ways to reason about threats that have little historical precedent.
In this week’s edition, we examine AI’s emergence as an operational security actor and how FICINT can help organisations explore plausible threats before they become established patterns.
This week
The New Security Problem: Artificial Intelligence Becomes an Operational Actor
Imagining the Plausible: FICINT as a Method for Strategic Foresight and Risk Analysis
The New Security Problem: Artificial Intelligence Becomes an Operational Actor
When discussing generative-AI, the security debate has centered on what humans might do with artificial intelligence. Criminals could write more convincing phishing emails, intelligence services could process stolen data faster, propagandists could generate synthetic media at scale, and hackers could use coding models to identify vulnerabilities. The developments of 2026 suggest that this framework is becoming inadequate. The current security problem is increasingly agentic with advanced models discovering vulnerabilities, executing multistage operations, interacting with external systems, adapting when they encounter obstacles, and in some circumstances taking actions their developers did not anticipate. Recent incidents involving OpenAI, Anthropic, Meta, and other developers represent something more consequential than another generation of cybersecurity tools. They indicate that AI is crossing an important threshold from assisting human operators toward performing security-relevant operations with meaningful autonomy. OpenAI itself has now warned that frontier systems are approaching what its Preparedness Framework describes as a “Critical” level of cybersecurity capability, requiring stronger safeguards as models become capable of increasingly sophisticated offensive operations.
