Safety brakes inside the labs (events, trailing year)

Deliberate slowdowns announced by a frontier lab and recorded as dated evidence rows with a verbatim announcement: paused training runs, gated releases, confinement of a model under a preparedness framework. Derived metric safety_brake_events_12m counts them over the trailing twelve months.

Why it matters. The paper's safety speed limits are argued from regulated domains; the tracker also watches for them on the invention stage, where a brake inside a lab slows the whole chain before any regulator acts.

Proxy types

policy, model_release

Unit

count

Cadence

quarterly

Valve

adoption to adaptation

• derived metric safety_brake_events_12m (formula in the semantic layer)

Three brakes in five weeks of 2026, all inside OpenAI and all self-announced; the count is a floor, since labs announce selectively.

3 observations. Hollow points are disputed (see counterevidence). Every point links to its observation.

1. 2026-09-06contextDt7grade D, evidence tier 7source

   Pachocki: voluntary slowdowns should become commonplace until shared safety bars exist (6 Sep 2026)

   “I expect and hope for voluntary slowdowns to become commonplace until shared safety bars are established.”

2. 2026-09-01supportsDt7grade D, evidence tier 7source

   Astra judged to meet the Critical cybersecurity threshold; release gated behind safeguards (1 Sep 2026)

   “We now believe Astra meets the Critical cybersecurity capability threshold under our Preparedness Framework, meaning that with the right tools and access, it can find previously unknown security flaws and develop ways to exploit them across many well-protected systems without a person guiding each step.”

3. 2026-08-18supportsDt7grade D, evidence tier 7source

   two-week pause in RL training on deployment-bound models after the Hugging Face incident (Jul-Aug 2026)

   “This included a two-week pause in reinforcement learning (RL) training on our latest models intended for deployment while we further hardened and red-teamed our research environments and expanded the coverage of our monitoring systems.”

4. 2026-08-07supportsDt7grade D, evidence tier 7source

   Astra internal activities paused pending strengthened security controls (7 Aug 2026)

   “We are pausing internal activities involving Astra that do not yet meet these strengthened security control requirements.”

Three announced brakes inside OpenAI in the trailing year, all with a verbatim source via the Internet Archive: internal Astra activity paused pending stronger security controls (7 Aug 2026), a two-week pause in RL training on deployment-bound models after the Hugging Face incident (18 Aug), and Astra judged to meet the Critical cybersecurity threshold with release gated behind safeguards (1 Sep). At or above the normal band (one or more a year), but every row is the lab describing itself (tier 7), so the status is capped at emerging. Initial seed.

• 2026-09-01 · 3as of 2026-09-01(3 obs)
• 2026-08-18 · 2as of 2026-08-18(2 obs)
• 2026-08-07 · 1as of 2026-08-07

1. 2026-09-10○unmeasured →to ◐emergingconf — → 40 · evaluate

   Three announced brakes inside OpenAI in the trailing year, all with a verbatim source via the Internet Archive: internal Astra activity paused pending stronger security controls (7 Aug 2026), a two-week pause in RL training on deployment-bound models after the Hugging Face incident (18 Aug), and Astra judged to meet the Critical cybersecurity threshold with release gated behind safeguards (1 Sep). At or above the normal band (one or more a year), but every row is the lab describing itself (tier 7), so the status is capped at emerging. Initial seed.

   obs:2ab8fec4obs:67d75927obs:82685d1e

40 / 95 — limited or vague evidence

Confidence is independent of status: 90–95 multiple strong independent sources; 70–89 good evidence, some ambiguity; 50–69 mixed or hard to operationalise; below 50 limited or vague.

• Human interventions on 4–8 hour agent tasks (self-reported) ◐emerging
• Agent-workdays per human workday in frontier research (self-reported) ◐emerging