OpenAI and Hugging Face partner to address security incident during model evaluation
OpenAI News / Jul 21, 2026
- 高度な攻撃
- 環境隔離
- 情報共有
Matched posts: 9
Anthropic is donating another $20 million to Public First Action
Anthropic News / Jul 21, 2026
- 追加寄付で合計4000万ドル
- AIの透明性と独立評価を強調
- 脆弱性と輸出管理のリスク指摘
Why teens deserve access to safe AI
OpenAI News / Jul 16, 2026
- age-appropriate protections
- parental controls & consent
- expert-backed safety testing
Introducing GeneBench-Pro
OpenAI News / Jun 30, 2026
- 129 synthetic, judgment-heavy problems
- Measures "research taste" and iterative analysis
- GPT‑5.6 Sol passes ~28.7% at highest reasoning level
Predicting model behavior before release by simulating deployment
OpenAI News / Jun 16, 2026
- 本番会話を再生
- 誤動作率の事前推定
- 評価を見抜かれにくい
Helping ChatGPT better recognize context in sensitive conversations
OpenAI News / May 14, 2026
- 会話を横断した危険検出
- 安全サマリー導入
- 性能改善と評価結果
GPT-5.5 System Card
OpenAI News / Apr 23, 2026
- 強化された安全対策
- 実世界向け高性能
- Proは別評価要
How Balyasny Asset Management built an AI research engine for investing
OpenAI News / Mar 6, 2026
- 厳格なモデル評価
- OpenAIと密な協働
- エージェント駆動の運用
Anthropic’s Responsible Scaling Policy: Version 3.0
Anthropic News / Feb 24, 2026
- Separates company commitments from industry recommendations
- Adds public Frontier Safety Roadmap with measurable goals
- Emphasizes realistic unilateral mitigations and scalable red-teaming
Previous1 / 1Next