OpenAI Strengthens GPT-6.2 Astra Against Reasoning Extraction
Unite.AI ·
Summary
OpenAI on November 20, 2026 announced stronger defenses against adversarial distillation, reporting that the average success rate for adaptive attacks extracting protected reasoning from an earlier request fell from 54.0% on GPT-6 Astra with original safeguards to 0.2% on GPT-6.2
- Source
- Unite.AI
- Published (IST)
Original reporting belongs to Unite.AI. DEEKOYA links the headline and a short excerpt only; we do not republish the full article.
Explore more: Today in AI · Research papers · AI tools · Learn AI
Related stories
Unite.AI
Thinking Machines Lab Releases Inkling 2 Mini for Local Devices
Thinking Machines Lab released Inkling 2 Mini on November 20, 2026, an open-weights model built to run on consumer…
Unite.AIModels
The Decoder
OpenAI makes it much harder for attackers to extract how its latest AI models think
OpenAI has significantly strengthened safeguards against extracting internal reasoning from GPT-6.2 Astra and GPT-6.2…
The DecoderModels
- Anadolu Ajansı
California proposes emergency powers to shut down high-risk AI models
Anadolu Ajansısafety
Unite.AI
METR Finds AI Time Horizons Doubling as Research Judgment Lags
METR reported on November 20, 2026, that public frontier models’ 50% time horizons—the human-estimated task duration at…
Unite.AIModels
AWS Machine Learning Blog
Announcing Amazon Textract Insights: Verifiable generative AI for documents
Announcing the general availability of Amazon Textract Insights, a new generative AI feature of Amazon Textract.…
AWS Machine Learning BlogResearch
Engadget
Signal now blocks screenshots by default to stymie AI agents
It describes the fix as incomplete and calls for a simple way for developers to opt out of AI snooping.
EngadgetAI
Who reported “OpenAI Strengthens GPT-6.2 Astra Against Reasoning Extraction”?
Unite.AI, on 21 Nov 2026. The full article is at the source link on this page.
Comments
No comments yet. Share your own experience below.
Visitor posts are moderated before they appear and are the opinion of their authors, not DEEKOYA AI. Names are not verified. See our privacy policy.