A Better Newspaper

## Overview A new report from threat intelligence firm Flashpoint finds that criminals have moved AI 'out of testing and into daily use,' well past the experimental phase (SiliconANGLE, August 13, 2026). The 2026 Global Threat Intelligence Report: Midyear Edition covers the first six months of the year and is based on analysis of 3.9 petabytes of material, most of it drawn from criminal sources (SiliconANGLE, August 13, 2026). ## Key Findings - Criminal groups reportedly now use AI as a standard operational tool rather than a novelty or experimental capability (SiliconANGLE, August 13, 2026). - The report is positioned as a systematic, large-scale analysis (3.9 petabytes) rather than anecdotal reporting, lending it credibility as an industry benchmark (SiliconANGLE, August 13, 2026). ## Why It Matters This finding fits into a broader 2026 narrative of AI-enabled offensive capability proliferation — including newly marketed 'cyber capable' frontier models (e.g., GLM-5.3), documented first cases of AI-assisted zero-day exploit development, and AI-enabled adaptive computer worms. For cybersecurity teams, insurers, regulators, and enterprises, the shift from experimental to operational criminal AI use has direct implications for risk modeling, security budgets, and policy responses (e.g., export controls, model access restrictions). ## Key Watch Points - Follow-on Flashpoint reports and corroborating findings from other threat intelligence firms. - Regulatory responses citing this data (e.g., in export control or AI safety policy debates). - Correlation with rising cyber insurance premiums and AI-specific coverage products (e.g., Cowbell Prime One).