China is set to introduce safety inspections of generative artificial intelligence services before they are launched. The regulations require AI service suppliers to ensure the accuracy of content and the protection of intellectual property without discrimination and compromised security. Additionally, AI operators must clearly label generated content.
The new rules will only apply to specific industries, according to the government’s Ministry of Science and Technology officials. This move comes after officials announced their intention to promote the safe and controlled application of AI in February, deeming it a strategic industry. Beijing emphasized the need for transparency in regards to data and algorithms used in training models, revealing a heightened attention towards controlling confidential and valuable information.
Under these guidelines, service providers must provide information on the origin, size, and type of data used in the training process. AI platforms must also share the basic algorithms and other technologies used. The regulations will apply to platforms like ChatGPT, which adhere to strict bans on disputable or unwelcomed content on the internet.
The Chinese government’s strict approach is also expected to encourage companies like Baidu and Sensetime to provide clearer rules for future services. The new regulations will prohibit deviant content without clarification and are expected to emphasize the importance of ensuring AI services are used in a safe and secure manner.