The AI Weaponization Paradox
3 of 8 filers across 2 sectors are flagging higher disclosed risk. Visible since 2025Q4, recently cooling. Reached 4 sectors at its 2025Q4 peak, now 2 sectors. Primarily a risk story (73%), with a strategic overlay (18%). Present-tense — companies describing what is happening now. Recent filings describe "proliferation of AI-enabled fraud and ransomware attacks that continue to be reported globally."
Companies are caught between leveraging AI for security defenses and facing accelerating cyberattack threats powered by the same AI capabilities, while emerging risks around algorithm bias and IP exposure complicate adoption and deployment.
DISTINCT NEW FILERS PER QUARTER
✦ WHAT THE DIFF CAUGHT
Rhetoric shifts from forward-looking threat modeling (BAC, GS) to present-tense capability deployment (AVGO) and emerging non-cyber AI risks (MCD, DE) that may constrain adoption.
REPRESENTATIVE SIGNAL FROM FILINGS
“proliferation of AI-enabled fraud and ransomware attacks that continue to be reported globally”
AI-enabled fraud and ransomware attacks continue to proliferate globally, presenting an escalating cybersecurity threat.
“AI lowers the entry barriers to plan and execute cyberattacks, enables more personalized and harder to detect social engineering”
Artificial intelligence lowers barriers to planning and executing cyberattacks and enables more sophisticated attacks.
DRIVERS