news pappernews papper

近日,OpenAI前安全总监、超级对齐团队负责人Jan Leike在宣布辞职后首次公开发声。在其声明中,Leike对OpenAI追求华而不实的产品提出了批评,并强调了AGI(人工通用智能)安全的重要性。

Leike表示,他在OpenAI的时间中,团队面临了一个越来越大的挑战:计算资源的不足。这使得安全研究的推进变得异常困难。他指出,OpenAI肩负着全人类的沉重责任,应当将更多的精力集中在准备下一代模型、提升安全性、监控、预防潜在风险、增强对抗性鲁棒性、超级对齐、保密性、社会影响等关键议题上。

他强调,尽管安全文化和流程至关重要,但在过去几年里,它们却让位于追求眼球效应的产品。Leike呼吁OpenAI应更清晰地认识到如何驾驭和控制比我们更智能的人工智能系统,这是OpenAI无法回避的挑战,也是全人类共同面对的课题。

Leike的这番言论,无疑为OpenAI乃至整个AI行业的未来发展提出了重要的思考。如何在创新与责任之间找到平衡,如何在追求技术突破的同时,确保AI的安全与合规,这不仅是OpenAI需要回答的问题,也是所有AI从业者需要共同面对的挑战。

英语如下:

**Headline:** **Controversy Erupts Over OpenAI’s Flashy Product as Former Security Director Criticizes忽视AI Safety**

Keywords: OpenAI criticism, insufficient resources, prioritize AGI safety

News Content: ### Former Head of Super Alignment Criticizes OpenAI for Neglecting AGI Safety, Calls for Enhanced Responsibility

Recently, Jan Leike, the former Security Director and Head of Super Alignment at OpenAI, spoke out publicly for the first time since announcing his resignation. In his statement, Leike criticized OpenAI’s pursuit of flashy products and emphasized the importance of AGI (Artificial General Intelligence) safety.

Leike said that during his time at OpenAI, the team faced an increasingly huge challenge: the lack of computing resources, which made it异常困难 to advance security research. He pointed out that OpenAI carries a heavy responsibility for all of humanity and should focus more on key issues such as preparing the next-generation model, enhancing safety, monitoring, preventing potential risks, enhancing adversarial robustness, super-alignment, confidentiality, and social impact.

He stressed that although security culture and processes are crucial, they have taken a back seat to products aimed at grabbing eyeballs in the past few years. Leike called on OpenAI to clearly recognize how to harness and control artificial intelligence systems that are smarter than us, a challenge that OpenAI cannot avoid and that confronts all of humanity.

Leike’s comments无疑为OpenAI乃至整个AI行业的未来发展提出了重要的思考.如何在创新与责任之间找到平衡,如何在追求技术突破的同时,确保AI的安全与合规,这不仅是OpenAI需要回答的问题,也是所有AI从业者需要共同面对的挑战.

【来源】https://mp.weixin.qq.com/s/nERwuSXyXa7Blrsw-uW0EA

Views: 1

发表回复

您的邮箱地址不会被公开。 必填项已用 * 标注