ChatGPT
artifactAI & Compute · Defense & Military-Industrial
OpenAI's AI program is now inside the Pentagon.
Who they are
ChatGPT, the AI program made by OpenAI.
What they do
On August 31, 2026 the Pentagon launched versions of ChatGPT and Grok for its 3 million staff. ChatGPT went down for hours on September 3 along with Claude and Grok.
How it works
It is OpenAI's main product, run by Sam Altman's company. DOGE used it to cancel 1,477 science grants.
Why it matters
It took a place inside the Pentagon after Anthropic was banned. An MIT study found people using it showed up to 55% less brain connectivity.
What's new (Aug–Oct 2026)
Aug 31, 2026 — The Pentagon launched versions of ChatGPT and Grok for 3 million civilian and military staff, TechCrunch reported. Fortune reported its GenAI.mil system counts 1.7 million users among the 3 million workforce. (techcrunch.com ; fortune.com)
Sept 3, 2026 — ChatGPT, Claude and Grok all went down on Thursday. OpenAI blamed a routing error starting about 7:43am PT, Anthropic had an infrastructure problem fixed by 16:16 UTC, and SpaceX's Memphis computing center had an outage. All were back by 12:38pm PT. (theregister.com ; 9to5google.com ; cnet.com ; theverge.com)
The engine's record — word for word
[Live pass Oct 6 2026] 2026-08-31 — Pentagon launched versions of ChatGPT and Grok for 3 million civilian and military personnel (TechCrunch); GenAI.mil counts 1.7 million users among the 3 million workforce (Fortune). (
techcrunch.com ;
fortune.com)
[Live pass Oct 6 2026] 2026-09-03 — ChatGPT, Claude and Grok outages Thursday 3 Sep: OpenAI cited routing error from about 7:43am PT; Anthropic infrastructure issue, restored 16:16 UTC; SpaceX Memphis compute center outage; all back by 12:38pm PT (9to5Google). (
theregister.com ;
9to5google.com ;
cnet.com ;
theverge.com)
Follow the trail
in the news with
OpenAI2026-09-03 — ChatGPT, Claude and Grok outages Thursday 3 Sep: OpenAI cited routing error from about 7:43am PT; Anthropic infrastructure issue, restored 16:16 UT
in the news with
Anthropic2026-09-03 — ChatGPT, Claude and Grok outages Thursday 3 Sep: OpenAI cited routing error from about 7:43am PT; Anthropic infrastructure issue, restored 16:16 UT
in the news with
SpaceX2026-09-03 — ChatGPT, Claude and Grok outages Thursday 3 Sep: OpenAI cited routing error from about 7:43am PT; Anthropic infrastructure issue, restored 16:16 UT
in the news with
Claude2026-09-03 — ChatGPT, Claude and Grok outages Thursday 3 Sep: OpenAI cited routing error from about 7:43am PT; Anthropic infrastructure issue, restored 16:16 UT
in the news with
Grok2026-08-31 — Pentagon launched versions of ChatGPT and Grok for 3 million civilian and military personnel (TechCrunch); GenAI.mil counts 1.7 million users among
cross-reference
Elon Musk**Report #86:** the Maximus Inc structural-insulation finding ($4M absorbed of $61B DOGE de-obligations) is the empirical inverse of the e_doge_war pattern (DOG
cross-reference
Feb 27 2026: Anthropic Banned2026-08-31 — Pentagon launched versions of ChatGPT and Grok for 3 million civilian and military personnel (TechCrunch); GenAI.mil counts 1.7 million users among
cross-reference
Mar 12 2026: DOGE War FailuresChatGPT used to cancel 1,477 science grants.
cross-reference
Sam AltmanRuns the engine's primary epistemic-capture surface (ChatGPT) plus biometric-ID infrastructure via Worldcoin (iris-scanning digital ID).
cross-reference
Hegseth Ultimatum (Feb 27 2026)2026-08-31 — Pentagon launched versions of ChatGPT and Grok for 3 million civilian and military personnel (TechCrunch); GenAI.mil counts 1.7 million users among
cross-reference
AI-Channeling Convergence (2024-2026)Documented emergence: 2024-2026 hybrid operators using ChatGPT / Claude / Gemini as 'spirit guide' / 'channeling intermediary' / 'AI oracle' interface.
cross-reference
Defense Innovation Board (DIB, 2016) — Schmidt-Chaired DoD-to-Silicon-Valley Wedge2026-08-31 — Pentagon launched versions of ChatGPT and Grok for 3 million civilian and military personnel (TechCrunch); GenAI.mil counts 1.7 million users among
cross-reference
AI Sexual-Content Carve-Out (OpenAI Model Spec: age-appropriate erotica/gore vs sexual-deepfake hard line)OpenAI Model Spec (2025) reclassifies generation of erotica and gore from prohibited to PERMITTED within 'age-appropriate' fictional contexts via API/ChatGPT, w
cross-reference
AI Cognitive Offloading — 'Cognitive Debt'MIT Media Lab 'Your Brain on ChatGPT' (Kosmyna et al., 2025/26; EEG, 54 subjects, 4 months): LLM users showed up to a 55% reduction in neural connectivity (alph
Walk this on the live map →