The founder of a leading Chinese artificial intelligence lab is pushing back against a growing instinct to lock up the most advanced AI systems.
According to Bloomberg, Tang Jie — the founder of the Chinese lab Zhipu, which operates under the name Z.ai and is behind the GLM family of AI models — argued in a memo that frontier AI capabilities should remain "as open and widely accessible as possible."
Bloomberg reports that Tang's position is that cutting-edge AI should stay broadly accessible rather than restricted. In practice, "open" access can mean releasing the underlying models so that outside developers, researchers, and companies can freely download, inspect, and build on them, instead of reaching the technology only through a paid, tightly controlled service.
The stance stands in contrast to the approach taken by some of the best-funded AI developers, who have kept their most powerful systems closed and proprietary, citing commercial and safety concerns. Z.ai's GLM models are among the more prominent AI efforts to emerge from China.
The source material here is limited to a single report, and it does not detail the full reasoning behind the memo or any specific policy or product commitments tied to it.
Why it matters: how the world's leading labs answer the open-versus-closed question will shape who gets to use, scrutinize, and profit from the most capable AI — and a Chinese lab publicly championing openness adds a notable voice to a debate that increasingly cuts across national and corporate lines.