Google Threat Intelligence reported on October 8 that vulnerability disclosures had doubled in its analysis and that AI-assisted discoveries were proportionally less concentrated in low-risk issues, with more moderate findings and remote-code-execution cases. It also reported an average of 11 zero-day exploitations per month from January through August 2026, versus eight per month in 2025.
This is a company threat research based on a primary source published October 8, 2026. AI Speed distinguishes a shipped feature or completed event from a preview, plan, study or vendor-selected case study. That distinction matters before a reader turns an announcement into a purchasing or policy decision.
What the evidence establishes
The article provides defined company research rather than a product promise. The numbers should be read with its collection method, taxonomy and coverage. A rise can reflect better discovery, disclosure practice or attacker activity, not one simple causal effect from AI.
Treat every number according to its source. Company metrics can be useful and specific without being independently verified. Record the product version, sample, baseline and measurement period; where those details are missing, describe the figure as a reported claim rather than a general result.
Related sources: Google Threat Intelligence: vulnerability trends in the AI era ↗
Limits and operational risks
Observed zero-days are an incomplete sample, and severity does not equal exploitability in every environment. AI-generated proofs of concept can be wrong or dangerous. Defensive use must remain authorized and isolated.
The full system includes identities, data, connectors, permissions, people and rollback—not only a model or cloud service. Test failure paths, prompt injection, stale information and cancellation. Keep a human accountable when output can affect money, access, employment, health, safety or publication.
Related sources: Google Threat Intelligence: vulnerability trends in the AI era ↗
What it means in Canada
Canadian organizations should shorten patch triage, inventory exposed assets and require evidence before emergency changes. Critical infrastructure teams need reporting and coordination paths that do not leak exploit details.
A bounded pilot should name its lawful purpose, data classes, processing locations, success threshold, owner and maximum cost. Canadian-region hosting can help but does not by itself settle privacy, sovereignty, accessibility or sector rules. Obtain qualified advice for high-consequence use.
What to watch next
Watch comparable datasets, time-to-patch, exploitation after disclosure and independent research on whether AI changes attacker success rates.
Recheck the linked primary source before acting because availability, pricing and documentation can change. A later rollout, benchmark or incident is new evidence; it should not be backdated into what was actually known on October 8.
Related sources: Google Threat Intelligence: vulnerability trends in the AI era ↗
Sources and editorial method
We use the linked original statements and reports to distinguish documented facts from our analysis. Company announcements describe their authors’ plans and claims; they are not independent verification of future outcomes. This article is dated and will be revised if material evidence changes. Read our editorial standards and corrections process.
- Google Threat Intelligence: vulnerability trends in the AI era ↗ — Primary source published October 8, 2026; accessed October 9, 2026. Classified as a company threat research; company claims remain attributed.