[1/3]An AI (artificial intelligence) billboard seen at the World Conference on Artificial Intelligence (WAIC) held in Shanghai, China, on July 6, 2023. Reuters/Aly Song
HONG KONG, July 12 (Reuters) – China also joins the global rush towards generative artificial intelligence, boasting nearly 80 AI models from companies such as Baidu (9888.HK) and Alibaba (9988.HK), and the start-up The company has raised nearly $14 billion. over the past 6 months.
However, unlike in the West, where OpenAI’s ChatGPT attracts more than 100 million monthly users and learns their input to continuously refine and improve their products, China’s AI chatbots are open to the public. not.
It speaks to a new reality in China’s tech sector and the effectiveness of the Chinese government’s years of regulatory crackdowns, which culminated in hefty fines against Ant Group and Tencent last week.
Once known for its fierce “market competition” spirit, especially in the consumer internet space, companies are taking a cue from Beijing and slowing down to defend their lines.
The formerly freewheeling industry is now seen as largely compliant, with companies from Alibaba to Tencent (0700.HK) restructuring their operations, slowing expansion into new areas and hiring thousands of workers. dismissing employees.
The new rules that Chinese companies must comply with range from scrutinizing algorithms to accepting security reviews for data they want to export.
But while the Chinese government now grasps what it sees as excesses and misbehavior stemming from the industry’s runaway growth, some experts say it’s a balance businesses must strike a strategy for. They are now sounding alarm bells about the implications for future competitiveness with Western rivals. .
Chinese regulators said last week that most of the problems with the country’s platform companies had been rectified, a move seen as a signal that the crackdown was finally over.
However, Henry Gao, a law professor at the Singapore Management University, said the tough regulatory environment, which is currently tougher in many areas than in other countries, is expected to continue.
“In recent years, there has been a great deal of what I call ‘preemptive regulations’ in China.
beta test
Industry executives say these challenges are growing as global interest in generative AI techniques rises following the success of Microsoft’s (MSFT.O)-backed ChatGPT, and as Chinese companies race to catch up to their U.S. peers. said to be particularly serious in
From Baidu’s Ernie Bot to SenseTime’s (0200.HK)’s SenseChat, there have been some big announcements and big announcements in China, but they’re all still in beta testing and open to select users only. It has been.
Baidu CEO Robin Lee said in May that he was waiting for government approval for Erniebot.
“The absence of a formal regulatory framework has prevented many Chinese internet companies from conducting large-scale testing of generative AI products that are in development or have already been developed,” said the head of China Internet stock research. Jialong Shi said. Mr. Nomura.
According to the chatbot, gaining more users will improve ChatGPT’s model to better consider language patterns such as slang and idioms, detect errors, handle unusual scenarios, and improve culture in responses. can reduce prejudice.
Mark Natkin, managing director of research firm Merbridge Consulting, said the main motivation for China’s reluctance to roll out AI chatbots is that uncensored chatbots change society’s perceptions in potentially disruptive ways. He said it was something the Chinese government feared it would start to affect.
“Press the brake pedal”
In April, China outlined a set of draft measures for generative AI services, saying companies would have to submit security assessments to authorities before launching their products.
Analysts said some of the rules were particularly onerous, such as the obligation for AI service providers to verify the truthfulness and accuracy of not only what their AI models produce, but also the material used to train them. criticize.
The measures are expected to be finalized and implemented by the end of this year. The Chinese government is also preparing an AI-specific law, but has not disclosed details.
Indeed, as countries around the world weigh safety and copyright while maintaining an environment conducive to innovation, and working to establish guardrails for hot tech, China appears to be ahead of regulation. .
But Heatherm Huang, co-founder of Hong Kong-based tech company Measureable AI, said this puts Chinese companies at a disadvantage because US products such as Google’s Bard and Microsoft’s Bing are being marketed straight out of the box. He expressed the view that it would be placed in
“It’s like putting a speed limit on a race car, it slows you down,” he says. “While the U.S. races ahead with AI, China is putting the brakes on more rules.”
business focus
Many of China’s tech companies are currently focused on finding industrial applications, a move that analysts say aligns with the Chinese government’s priorities and a key goal of the new regulatory framework. said.
Chinese President Xi Jinping has stressed that he wants to put more energy into “hard” technologies such as semiconductors and industrial AI applications, as well as breakthrough technologies that help China reduce its dependence on Western technology.
Huawei Technologies (HWT.UL) said last week that its Pangu AI model will primarily serve enterprise customers in applications such as freight train safety inspections and weather forecasting.
Baidu announced last week that more than 150,000 businesses have applied to test ErnieBot since its launch in March, with more than 300 testing in areas such as improving office efficiency, customer service and marketing. .
Reported by Josh Ye.Editing: Brenda Goh and Jamie Freed
Our standards: Thomson Reuters Trust Principles.
