China moves to regulate ChatGPT-like AI chatbots

The Chinese government is moving to regulate the nation’s burgeoning chatbot industry, taking a cautious approach to the technology to maintain tight control over its development. The draft rules unveiled by the Cyberspace Administration of China (CAC) this month, which will affect artificial intelligence (AI) systems like ChatGPT used to generate text and images in response to user requests, demonstrate Beijing’s resolve to control the burgeoning technology sector.

Under the new regulations, AI companies must adhere to the Chinese Communist Party’s censorship guidelines, avoiding materials that tarnish China’s leaders or delve into prohibited history. Chatbots are expected to promote “socialist core values” while shunning information that threatens “state power” and national unity.

In addition, the draft rules stipulate that chatbot creators must ensure their products produce truthful content that protects intellectual property. Companies are required to register their algorithms, the software brains behind chatbots, with the appropriate regulators.

Although the regulations are not yet final and may undergo further revisions, AI engineers in China are already contemplating how to integrate Beijing’s edicts into their products.

The ChatGPT phenomenon has fascinated governments around the world. Developed by the US-based OpenAI, which boasts around US$13 billion in backing from Microsoft, the AI has prompted Silicon Valley to explore applications in industries such as video games and advertising. According to venture capital firm Sequoia Capital, AI businesses could eventually generate “trillions of dollars” in economic value.

As a result, China’s investors and entrepreneurs are proceeding at breakneck speeds to keep up. Shares in Chinese AI firms have surged, with e-commerce giant Alibaba, facial recognition creator SenseTime, and search engine Baidu among those making waves in the space. Furthermore, at least two startups developing Chinese rivals to OpenAI’s technology have raised millions of dollars.

Since ChatGPT is not accessible in China, the country has swiftly established its red lines for AI ahead of other nations pondering regulation of chatbot technology. Some experts suggest China’s rules demand a level of technical control over chatbots unattained by its tech firms.

Analysts expect the new rules to be refined following consultations with China’s tech companies, and enforcement may be softened so that they do not hinder technological progress.

China has a well-established practice of monitoring the Internet. Over the past two decades, it created the largest information dragnet in the world, scaring off Western tech companies that did not wish to comply with its regulations. Meanwhile, China’s local tech firms thrived by adhering to the rules.

However, the rise of AI chatbots presents a new challenge to the Communist Party. The unpredictable nature of generative AI, which has the potential to create false or nonsensical statements, contradicts the party’s mission of information control.

Chatbot development and AI technology in China are expected to continue despite the country’s struggles with semiconductor competition and sanctions. For example, ChatYuan, a Chinese version of ChatGPT, released in February, was downloaded thousands of times before being subsequently removed from the market over controversial responses.

The app’s creator, AI engineer and entrepreneur Xu Liang, is reportedly working on a more “patriotic” bot, incorporating keyword filtering and additional manual review processes. This evolution could mark a new chapter in the regulation and orientation of China’s AI chatbot industry.