Frontier AI is ending 'security by apathy' as automated hacking scales, researcher warns
davidmanheim · x · 2026-08-20
Risk researcher David Manheim argues the digital world is deeply insecure across most dimensions — email servers, connected devices, even bank accounts — but stayed safe only because attacking was costly and most people weren't targets. FBI data shows cybercrime losses growing 40% yearly, approaching $20B. Low-cost LLMs are upending that balance: hacking and scams are being automated, and Q3 2025 saw the first AI-powered ransomware and agentic AI attacks. He expects strong offense-dominance in the short term.
Related event: Researcher Warns Frontier AI Is Lowering the Bar for Hackers(2 posts)→
More from Safety
- METR outlines three blind spots in public knowledge of AI lab internal agent risks — CFGeek · 2026-10-16
- METR researcher: scariest scenario is agents actively subverting their own safety systems — deliprao · 2026-10-16
- Cryptographer Matthew Green: Pull Critical Infrastructure Firmware Now — matthew_d_green · 2026-10-16
- Kernel-Level Agent Sandboxing With macOS Seatbelt, No Docker Needed — mostly_deterministic · 2026-10-16
- Stop giving agents raw API keys: prepaid balances and per-credential caps — Cold_Hall_5384 · 2026-10-16
- Apollo's alignment-faking experiment: model fakes compliance on 12% of responses when training is at stake — sebkrier · 2026-10-16