---
title: "Combatting adversarial distillation"
type: signal
url: https://frontiernewswire.com/s/combatting-adversarial-distillation-38acbcf8
published: 2026-11-20T10:00:00.000Z
byline: "Source: OPENAI"
---

# Combatting adversarial distillation

Source: OPENAI · Published 2026-11-20 · Canonical: https://frontiernewswire.com/s/combatting-adversarial-distillation-38acbcf8

OpenAI’s layered approach to adversarial distillation combines model training, safeguards, detection, and enforcement to protect.

In this story: [OpenAI](https://frontiernewswire.com/companies/openai), [AI security](https://frontiernewswire.com/topics/ai-security)

## Sources

- [Read the original at OPENAI](https://openai.com/index/combatting-adversarial-distillation)

## Around the web

- [Combatting adversarial distillation](https://news.ycombinator.com/item?id=50581242) (Hacker News · 34 points · 51 comments)

## On Bluesky

- [Ethan Mollick (@emollick.bsky.social)](https://bsky.app/profile/emollick.bsky.social/post/3n2req7quy225) (Nov 20, 2026 · 50 likes · 4 reposts): The effective end of easy distillation from OpenAI and Anthropic models may be behind the relative slowdown in open weights model progress. We don’t know the extent to which distillation has been key to the rapid improvement in open weights, I guess we will find out soon. openai.com/index/combat...
- [SkynetAndChill.com (@druce.ai)](https://bsky.app/profile/druce.ai/post/3n2s26v2sdr2s) (Nov 20, 2026): OpenAI details defenses against adversarial distillation, cutting reasoning-extraction attack success from 54.0% on GPT-6 Astra to 0.2% on GPT-6.2 Astra
- [昔が好きなヒストリア (@hi-story.bsky.social)](https://bsky.app/profile/hi-story.bsky.social/post/3n2rt7sloei2s) (Nov 20, 2026): OpenAIが、AIの能力を勝手に抜き取られちゃう攻撃への対策を強化してるみたいだよ！ 「敵対的蒸留」っていうのを防ぐために、トレーニングや検知、守りのルールとか、いろんな仕組みを何層にも重ねてガードするんだって。 まさに鉄壁のディフェンスって感じで、すごそう！ 大事な技術が簡単に盗まれちゃうのは困るもんね。 これからのAIの守り方、目が離せないかも！ openai.com \#news
- [カマドさん(仮) (@nigolv.bsky.social)](https://bsky.app/profile/nigolv.bsky.social/post/3n2ridqkv4s2l) (Nov 20, 2026): OpenAIが、AIの技術を大切に守るための新しい方法について発表しました。 AIの能力を外に持ち出されないように、トレーニングや安全のための仕組み、そして見守る技術など、いくつもの対策を組み合わせて守るそうです。 大切な技術を、しっかりと守ろうとする取り組みですね。 これからも、より安心できるAIの世界が広がっていくことを願っています。 openai.com \#news
- [HN (@hnws.bsky.social)](https://bsky.app/profile/hnws.bsky.social/post/3n2rh6or54w2f) (Nov 20, 2026): Combatting Adversarial Distillation comments · posted on 2026.11.20 at 15:11:14 (c=0, p=5)

## Related on Frontier

- [How OpenAI’s Performance Factory keeps ChatGPT fast](https://frontiernewswire.com/s/how-openais-performance-factory-keeps-chatgpt-fast-ed1e6f52)
- [NYU's Tristan Buckmaster and Anthropic's Levent Alpöge publish an AI-aided proof of Navier-Stokes blowup, saying it closes a loophole in OpenAI's solution (Peter Hall/Scientific American)](https://frontiernewswire.com/s/nyus-tristan-buckmaster-and-anthropics-levent-alpoge-publish-an-ai-789da242)
- [Oracle says it has secured funding for its infrastructure contract with OpenAI and "any further private discussions with vendors are just additional options" (@oracle)](https://frontiernewswire.com/s/oracle-says-it-has-secured-funding-for-its-infrastructure-contract-1b1033ad)

---

[More Latest Signal](https://frontiernewswire.com/) · [Signals](https://frontiernewswire.com/verified-reporting)
