Question 1

欧盟委员会关于 OpenAI 和 Anthropic 宣布了哪些内容？

Accepted Answer

Commission spokesperson Thomas Regnier confirmed on May 11 that the EU is holding separate discussions with both companies. OpenAI proactively offered access to its new AI model, while Anthropic has held four or five meetings with officials but hasn't gotten to the topic of model access yet.

Question 2

为何欧盟现在直接介入与 AI 开发者的对话？

Accepted Answer

The talks come as the EU enforces its AI Act and governments globally ramp up scrutiny of frontier AI systems. Recent findings showed that models from both OpenAI and Anthropic demonstrated advanced cybersecurity capabilities, including the ability to exploit software vulnerabilities, which raised safety concerns among regulators.

Question 3

欧盟的方法与美国正在考虑的方案有何不同？

Accepted Answer

The EU is pursuing voluntary discussions with AI companies under its risk-based AI Act framework. The White House, by contrast, is weighing a more direct intervention through mandatory pre-release government review of the most powerful AI models before they reach the public. The EU's Act doesn't currently require centralized pre-release review, making its approach less interventionist than what Washington is considering.

《人工智能法案》

欧盟表示，《人工智能法案》使其“具备充分能力”来遏制失控的AI代理

一份犀利简报。
每天。

《人工智能法案》

欧盟表示，《人工智能法案》使其“具备充分能力”来遏制失控的AI代理

一份犀利简报。每天。

一份犀利简报。
每天。