Zhipu Founder Advocates for Open Access to Frontier AI
Zhipu's founder, Tang Jie, has argued that frontier AI should remain accessible to everyone, contradicting China's potential plans to limit overseas access to its advanced open models. This stance puts him at odds with his government, which is considering restrictions on open models.
Key points
- Zhipu's founder, Tang Jie, made the case for openness in an internal memo, arguing that broad participation and sharing are key to AI safety.
- Zhipu has released its GLM-5.2 model under an open-source license, free to download and commercialize.
- China's government is considering limits on overseas access to its most advanced open models, contradicting Zhipu's stance.
- Zhipu's models have gained global popularity due to their free and open nature, with Chinese models now competing with US labs.
- The EU has not commented on the issue, but analysts say it highlights the tension between openness and security in AI development.
The debate over openness in AI development has taken a new turn with Zhipu's founder, Tang Jie, advocating for continued access to frontier AI. In an internal memo, Tang argued that broad participation and sharing are essential for AI safety, rather than relying on technological barriers.
Zhipu has already put its words into action, releasing its GLM-5.2 model under an open-source license. This allows users to download and commercialize the model, further increasing its global popularity. However, this move puts Tang at odds with his government, which is considering limits on overseas access to its most advanced open models.
The potential restrictions are a significant shift in China's stance on AI openness. For years, China has benefited from its open-source approach, with its models gaining traction globally. Now, Chinese models are competing with US labs, and some officials are questioning whether the country has given away too much.
The EU has not commented on the issue, but analysts say it highlights the tension between openness and security in AI development. As the debate continues, one thing is clear: the future of AI access will be shaped by the balance between these competing interests.
Sources
The WireByte editorial team synthesises technology news from multiple primary sources, verifies the facts, and links every source. Articles are produced with AI assistance and reviewed under our editorial policy.