diff --git a/entities/ai-alignment/anthropic.md b/entities/ai-alignment/anthropic.md index 91ffcf84a..e5c9f134e 100644 --- a/entities/ai-alignment/anthropic.md +++ b/entities/ai-alignment/anthropic.md @@ -68,6 +68,9 @@ Frontier AI safety laboratory founded by former OpenAI VP of Research Dario Amod - **2025-05-01** — Activated ASL-3 protections for Claude Opus 4 as precautionary measure without confirmed threshold crossing, first model that could not be positively ruled below ASL-3 thresholds - **2025-05-01** — Activated ASL-3 protections for Claude Opus 4 as precautionary measure without confirmed threshold crossing, first model that could not be positively ruled out as below ASL-3 capability levels - **2025-08-01** — Published persona vectors research demonstrating activation-based monitoring of behavioral traits (sycophancy, hallucination) in small open-source models, with 'preventative steering during training' reducing trait acquisition without capability loss +- **2026-02-15** — Pentagon sets February 27 deadline for Anthropic to comply with 'any lawful use' requirement, threatening contract termination and national security penalties +- **2026-02-24** — CEO Dario Amodei publicly refuses DoD demand, stating Anthropic cannot 'in good conscience' grant any-lawful-use authority for autonomous targeting and mass surveillance +- **2026-02-27** — Designated as supply chain risk by Trump administration, effectively blacklisting the company from Pentagon contracts due to hard red lines on autonomous weapons and mass surveillance. ## Competitive Position Strongest position in enterprise AI and coding. Revenue growth (10x YoY) outpaces all competitors. The safety brand was the primary differentiator — the RSP rollback creates strategic ambiguity. CEO publicly uncomfortable with power concentration while racing to concentrate it. diff --git a/entities/ai-alignment/openai.md b/entities/ai-alignment/openai.md index 4bff74f87..e2212429c 100644 --- a/entities/ai-alignment/openai.md +++ b/entities/ai-alignment/openai.md @@ -51,6 +51,7 @@ The largest and most-valued AI laboratory. OpenAI pioneered the transformer-base - **2026-03** — Released GPT-5/5.2/5.3 - **2026-03** — Restructured to Public Benefit Corporation - **2026-03** — IPO expected H2 2026-2027 +- **2026-02-28** — Announced Pentagon deal allowing military use of OpenAI technology under 'any lawful purpose' language with aspirational constraints on autonomous weapons and domestic surveillance, hours after Anthropic blacklisting. CEO Sam Altman described initial rollout as 'opportunistic and sloppy.' Amended March 2, 2026 to add 'intentionally' qualifier and exclude non-US persons from surveillance protections. ## Competitive Position Highest valuation and strongest consumer brand, but losing enterprise share to Anthropic. The Microsoft partnership (exclusive API hosting) provides distribution but also dependency. Key vulnerability: the enterprise coding market — where Anthropic's Claude Code dominates — may prove more valuable than consumer chat. diff --git a/inbox/queue/.extraction-debug/2026-02-28-govai-rsp-v3-analysis.json b/inbox/queue/.extraction-debug/2026-02-28-govai-rsp-v3-analysis.json new file mode 100644 index 000000000..1136b6747 --- /dev/null +++ b/inbox/queue/.extraction-debug/2026-02-28-govai-rsp-v3-analysis.json @@ -0,0 +1,27 @@ +{ + "rejected_claims": [ + { + "filename": "transparency-infrastructure-without-binding-commitments-creates-accountability-theater-not-safety-governance.md", + "issues": [ + "missing_attribution_extractor" + ] + } + ], + "validation_stats": { + "total": 1, + "kept": 0, + "fixed": 4, + "rejected": 1, + "fixes_applied": [ + "transparency-infrastructure-without-binding-commitments-creates-accountability-theater-not-safety-governance.md:set_created:2026-03-28", + "transparency-infrastructure-without-binding-commitments-creates-accountability-theater-not-safety-governance.md:stripped_wiki_link:voluntary-safety-pledges-cannot-survive-competitive-pressure", + "transparency-infrastructure-without-binding-commitments-creates-accountability-theater-not-safety-governance.md:stripped_wiki_link:only-binding-regulation-with-enforcement-teeth-changes-front", + "transparency-infrastructure-without-binding-commitments-creates-accountability-theater-not-safety-governance.md:stripped_wiki_link:AI-transparency-is-declining-not-improving-because-Stanford-" + ], + "rejections": [ + "transparency-infrastructure-without-binding-commitments-creates-accountability-theater-not-safety-governance.md:missing_attribution_extractor" + ] + }, + "model": "anthropic/claude-sonnet-4.5", + "date": "2026-03-28" +} \ No newline at end of file diff --git a/inbox/queue/2026-02-28-govai-rsp-v3-analysis.md b/inbox/queue/2026-02-28-govai-rsp-v3-analysis.md index d320e9de4..f2d199554 100644 --- a/inbox/queue/2026-02-28-govai-rsp-v3-analysis.md +++ b/inbox/queue/2026-02-28-govai-rsp-v3-analysis.md @@ -7,9 +7,12 @@ date: 2026-02-28 domain: ai-alignment secondary_domains: [] format: article -status: unprocessed +status: enrichment priority: medium tags: [RSP-v3, GovAI, responsible-scaling-policy, binding-commitments, pause-commitment, RAND-SL4, cyber-operations, CBRN, governance-analysis, weakening] +processed_by: theseus +processed_date: 2026-03-28 +extraction_model: "anthropic/claude-sonnet-4.5" --- ## Content @@ -57,3 +60,13 @@ GovAI's systematic analysis of what changed between RSP v2.2 and RSP v3.0 (effec PRIMARY CONNECTION: voluntary-pledges-fail-under-competition — pause removal is the clearest evidence; transparency-binding tradeoff is the new governance pattern to track WHY ARCHIVED: GovAI's analysis is the authoritative RSP v3.0 change log; the cyber/CBRN removal without explanation is the key unexplained governance fact EXTRACTION HINT: Focus on the transparency-without-binding-constraints pattern as a new KB claim — RSP v3.0 increases public accountability infrastructure (roadmaps, reports) while decreasing binding safety obligations, making it a test case for whether transparency without enforcement produces safety outcomes. + + +## Key Facts +- RSP v3.0 became effective February 24, 2026 +- GovAI published their analysis on February 28, 2026 +- RSP v3.0 requires interpretability-informed alignment assessment by October 2026 +- Frontier Safety Roadmap updates required approximately quarterly +- Risk Reports required every 3-6 months +- RAND Security Level 4 protections moved from binding commitment to industry-wide recommendation +- Cyber operations and radiological/nuclear removed from binding commitments without explanation