However, 59 per cent of EU respondents and 58 per cent of US respondents indicated they would only use such AI models hosted on non-Chinese infrastructure. That preference highlights the scrutiny Chinese models face globally, a concern that analysts suggested were not unfounded.
Chinese AI systems are often trained to provide “correct” answers regarding politically sensitive topics, a trait that may be viewed unfavourably by users in other countries, according to James Wang, general partner at US-based venture capital fund Creative Ventures.
03:34
Nvidia CEO Jensen Huang praises China’s AI progress following chip sales approval
Nvidia CEO Jensen Huang praises China’s AI progress following chip sales approval
“There are many specific historical or cultural questions that will lean towards either a Chinese context … or are deliberately reinforced to give the ‘correct’ answer from the Chinese government’s perspective,” Wang said.
