Why AI cyber capabilities surged since 2022 but cyber damages didn't: Joshua Saxe analysis
joshua_saxe · x · 2026-10-28
Security researcher Joshua Saxe published a long-form analysis with open data sources and statistical models probing a core puzzle: AI cyber capabilities (social engineering, vuln discovery, exploit writing, agentic hacking) grew rapidly from 2022-2026, yet total cyber damages didn't. Key findings: damages grew 3x over the past decade due to industrialized scams, fraud and ransomware; growth happened before ChatGPT/Llama 3 and actually tapered in the AI era; events like PHP2Shell were largely found and patched by AI-assisted defenders. He argues the non-response of damages to rising AI capability is the conceptual crux any AI security policy proposal must explain. Research assisted by Claude Opus 5.5, framed as a sketch toward a peer-review-quality paper.
More from Safety
- Webinar to present empirical research agenda for studying inter-AI morality — dioscuri · 2026-10-28
- New York State Court System Issues Official Guidance to Prevent AI Hallucinations — news-10 · 2026-10-28
- European 'sovereign AI' startup Giotto bans benchmarking in its ToS, drawing flak — teortaxesTex · 2026-10-28
- p(doom) became a popular term — how seriously should we take the risks? — Unfair-Journalist-66 · 2026-10-28
- EU governments approve GDPR change letting AI train on personal data without consent — IgorCarron · 2026-10-28
- Cyphrex Launches MCP Gateway to Block AI Agent Actions with Audit Trail — modelcontextprotocol · 2026-10-28