HONG KONG (Reuters) – China has joined the global boom in generative intelligence, boasting about 80 AI models from companies such as Baidu and Alibaba, and startups attracting nearly $14 billion in funding over the past six months.
But unlike in the West, where OpenAI’s ChatGPT attracts more than 100 million users every month, whose inputs are regularly reviewed to help improve the product, no Chinese AI chatbots have been made available to the general public.
It focused on the new reality of China’s tech sector and the effectiveness of years of Beijing’s regulatory crackdown, including last week’s hefty fines against Ant Group and Tencent.
Once characterized by a determined, “market competition” spirit, particularly in the consumer Internet sector, companies are slow to toe the line and take cues from Beijing.
The once free-wheeling industry now appears largely compliant, with companies from Alibaba to Tencent scaling back their businesses, slowing expansion into new areas and laying off thousands of workers.
Chinese companies must comply with the new rules, from vetting algorithms to receiving security reviews for exports.
But as Beijing now clamps down on what it sees as excesses and misbehavior stemming from the industry’s exodus, some experts are sounding the alarm about the impact on balanced companies’ future competitiveness with Western rivals. .
Chinese regulators announced last week that problems with most of the country’s platform companies have been fixed, a sign that the crackdown is finally over.
But Henry Gao, a law professor at Singapore Management University, said the tightly regulated environment, now tighter in many areas than in other countries, is expected to continue.
“In recent years, there have been a lot of what I call ‘pre-regulations’ in China. They definitely stifle innovation and slow down the ability of Chinese companies to catch up.”
Industry executives say these challenges are particularly acute in generative AI, as interest in the technology grows as Microsoft-backed ChatGPTT and Chinese companies race to catch up with their American peers.
There are many well-known startups in China, from Baidu’s Ernie Bot to SenseTime’s SenseChat, all of which remain in beta testing and open only to select users.
In May, Baidu CEO Robin Li told Ernie Bot that it was awaiting government approval.
Jialong Shi, head of China Internet Equity Research, said, “Because a common regulatory framework has not been established, many Chinese Internet companies cannot conduct large-scale testing of their AI products that are in development or have already been developed.” Nomura
Having more users will help ChatGPT refine its model to better capture language patterns like words and idioms, detect errors, respond to unusual situations and reduce cultural bias in responses, the chatbot said.
A key reason behind China’s desire to release AI chatbots is Beijing’s concern that uncensored chatbots could influence public views in ways that could influence them, said Mark Natkin, managing director of research firm Marbridge Consulting.
‘hitting the brakes’
China outlined draft measures for generative AI services in April, requiring companies to submit safety assessments to authorities before launching their products.
Analysts have criticized some of the rules as particularly harsh, such as requiring AI service providers to ensure the veracity and accuracy of not only what AI models produce, but also what is used to train them.
The moves are expected to be completed by the end of this year. The Chinese government is also preparing a standalone AI law, but has not provided details.
To be sure, China appears to be ahead of the regulatory curve as they establish safeguards against wildly popular technology around the world, balancing security and copyright to maintain an environment conducive to innovation.
But Heather Huang, founder of Hong Kong-based Measurable AI technology company, said she has seen American products like Google’s Bard and Microsoft’s Bing put Chinese companies at risk as they go directly to market.
“It’s like a speed limit on a race car — it slows things down,” he said. “While the US races ahead with AI, China is hitting the brakes with more regulations.”
Most Chinese tech companies are focused on finding applications for industrial use, a direction analysts say is in line with Beijing’s priorities and a key goal of the new regulatory framework.
Chinese President Xi Jinping has emphasized that he wants to put more power into “strong” technologies, such as semiconductors and industrial AI applications, and breakthroughs that will help China reduce its dependence on Western technology.
Huawei Technologies said last week that the Pangu AI model will be used mostly by enterprise customers for applications such as safety inspections of freight trains and weather forecasting.
Baidu said last week that more than 150,000 companies have applied to test Ernie Bot since it launched in March, with more than 300 testing it in areas such as office efficiency, customer service and marketing.
(Reporting by Josh Ye; Editing by Brenda Goh and Jamie Fried)