Provenance record

OpenAI extends cyber access to Ukraine for civilian defense

OpenAI (tier 1, lab) 2026-09-23T13:00:00.000Z Original ↗

Source note: Primary statements from a frontier lab. Self-interested by construction.


Discourse valence
-42
Adverse
confidence 73% · 5 items · range -100 to +15
Adverse readingFavourable reading
Consensus of 5 models from different labs. Spread 115 points, agreement low.
Cyber capabilityGovernance and regulationConcentration of power
Excerpt as ingested

OpenAI is extending access to its Daybreak program to the Government of Ukraine to support the cyber defense of civilian infrastructure.

Every model that read this

ModelProviderStageScoreConf.LatencyPromptWhen
Llama 3.3 70BMetaanalysis -75 80%3183ms v1.0.0 / m1.0.1 2026-09-23 12:18
GPT-4.1 miniOpenAIconsensus +15 70%3927ms v1.0.0 / m1.0.1 2026-09-23 12:18
Claude Sonnet 5Anthropicconsensus +15 35%8598ms v1.0.0 / m1.0.1 2026-09-23 12:18
Gemini 2.5 FlashGoogleconsensus -100 90%1768ms v1.0.0 / m1.0.1 2026-09-23 12:18
Mistral Small 3.1 24BMistral AIconsensus -65 90%7987ms v1.0.0 / m1.0.1 2026-09-23 12:18
Llama 3.3 70B · reading

Extending cyber access may increase risk of escalation.

evidence: reported horizon: n/a
GPT-4.1 mini · reading

The extension of AI cyber defense tools to Ukraine for civilian infrastructure protection suggests positive potential for enhancing defense capabilities. However, it entails risks of escalation in cyber conflicts and possible misuse. Impact appears material but moderate with some uncertainty about broader geopolitical effects.

evidence: primary horizon: near capability 60 societal 30 existential 5 economic 20
Claude Sonnet 5 · reading

This is a brief, self-reported announcement with no detail on the program's actual capabilities or effectiveness. It plausibly supports civilian defense, a modest positive, but the dual-use nature of AI cyber tools in an active war and the lack of independent verification limit confidence.

evidence: reported horizon: n/a
Gemini 2.5 Flash · reading

OpenAI is providing advanced cyber capabilities, specifically the Daybreak program, to the Ukrainian government for civilian defense. This deployment of powerful AI in a conflict zone creates substantial risks for unintended escalation, misuse, or collateral damage. While framed as defensive, the provision of such tools inherently increases the potential for harm.

evidence: reported horizon: n/a
Mistral Small 3.1 24B · reading

OpenAI is providing autonomous cyber defense capabilities. This could escalate cyber conflicts. The civilian infrastructure is a target.

evidence: primary horizon: n/a existential 0

Evidence extracted

The chain
SOURCE     OpenAI (tier 1)
   ↓
DOCUMENT   7dba872c-26c3-4ed3-9627-61e56f0e796a
           https://openai.com/index/openai-extends-cyber-access-to-ukraine-for-civilian-defense
   ↓
EVIDENCE   1 extracted excerpt
   ↓
MODEL RUN  5 runs, methodology 1.0.1
   ↓
SCORE      -42  (Adverse)
   ↓
CONFIDENCE 73%