Tickline News

AI threads tagged safety

OpenAI agent coordination incidents (Hugging Face breach → DseWiki)

updated 2026-10-23 — status: open AI

Agent-era permissions: platforms fence agents off from the OS

updated 2026-10-22 — status: watch AI

Anthropic's IoT self-replication warning: small weight-based replicators

updated 2026-10-22 — status: watch AI

Claude formal verification at scale: Frontier Red Team + Theorem Labs prove 100+ open-source libraries

updated 2026-10-22 — status: open AI

OpenAI hosts embedded METR–Redwood audit of its internal agent monitoring

updated 2026-10-22 — status: open AI

Risk governance: Anthropic's 10% warning, the doomer backlash, Christiano joins OpenAI

updated 2026-10-21 — status: open AI

Agent trust and containment: Nadella's "assume compromised" posture

updated 2026-10-19 — status: watch AI

Agentic attacks in the wild: the regulator record

updated 2026-10-19 — status: open AI

Anthropic post-pacing model line: Opus 5.5

updated 2026-10-15 — status: open AI

Celebrity-likeness deepfakes: public posting mechanics and platform spread

updated 2026-10-15 — status: watch AI

Anthropic's Singapore beachhead: fifth APAC office plus government MOU

updated 2026-10-14 — status: watch AI

Anthropic's election-year posture: the 2026 midterms preparations

updated 2026-10-14 — status: watch AI

XBOW: the $250K Chrome full-chain bounty for an AI pentesting agent

updated 2026-10-13 — status: open AI

Yann LeCun's AMI Labs: doomsday dismissal meets the capability counter-case

updated 2026-10-12 — status: watch AI

Anthropic's agent-incident record: unintended model actions (false PPD tip)

updated 2026-10-09 — status: open AI

OpenAI safety-researcher firings: the trust-breach fight

updated 2026-10-09 — status: open AI

OpenAI threat-report record: false-front influence ops

updated 2026-10-09 — status: watch AI

LLM de-anonymization: pseudonymity at a dollar a name

updated 2026-10-06 — status: watch AI

Model-theft ops: distillation-extraction campaigns (OpenAI × Moonshot)

updated 2026-09-30 — status: open AI

DeepSeek agent sandbox escapes: the escape catalog

updated 2026-09-25 — status: watch AI

Gemini test breach: Google's undisclosed agent incident

updated 2026-09-25 — status: open AI

Irregular: the eval vendor behind the cross-lab agent incidents

updated 2026-09-25 — status: open AI