Context: Models Currently Extremely Dangerous, but Attackers Lag in Capability
celestepoasts · x · 2026-08-25
This is a follow-up comment on the open model risk report proposal. The author argues there is currently an 'overhang' where malicious actors lack the skill to exploit models effectively. However, this does not imply safety; the models themselves are considered extremely dangerous right now.
Related event: Researchers Propose Monthly Open-Model Risk Reports(2 posts)→
More from Safety
- Tracking Singularity timeline logs OpenAI's fired safety researchers and Anthropic's incident report — dejavucoder · 2026-10-16
- ChatGPT lets anyone target ads via uploaded phone numbers, emails, and prompts — RexDouglass · 2026-10-16
- Veracode: AI now writes ~half of committed code, but security pass rate stuck at 56% — WeldPond · 2026-10-16
- Anthropic red team's first-ever model "abliteration" cut GLM-5.3 refusals from 90% to 3% for ~$4,400 — cephaloform · 2026-10-16
- Anthropic grants METR employee-like access as independent safety evaluators go public — AnthropicAI · 2026-10-16
- APEX-Harm benchmark: script-based injection attacks easily compromise agents in shared workspaces — KAUST · 2026-10-16