Five months after ChatGPT sparked an artificial intelligence investment frenzy, Beijing is moving to curb chatbots in China.
China’s Cyberspace Administration this month released draft rules for so-called generative artificial intelligence (the software system behind ChatGPT). This software system can create text and images in response to user questions and prompts.
According to the regulations, businesses must pay attention to the Chinese Communist Party’s strict censorship rules, ensuring that their websites and apps don’t publish material that defames China’s leaders or rehashes banned history. need to do it. The content of the AI system should reflect “core socialist values” and avoid information that undermines “state power” and national unity.
Companies also have to ensure that the words and images their chatbots create are truthful and respectful of intellectual property, registering the algorithms — the software brains behind chatbots — with regulators. need to do it.
The rules aren’t final and regulators may continue to amend them, but experts say engineers building artificial intelligence services in China have already figured out how to build orders into products. I’m here.
Governments around the world have been amazed by the power of chatbots, with AI-generated results ranging from surprising to harmless. Artificial intelligence is being used to create fake photos of Pope Francis, who passed college exams and wore a puffy coat.
Developed by US company OpenAI, which is backed by approximately $13 billion from Microsoft, ChatGPT has inspired Silicon Valley to apply its underlying technology to new areas such as video games and advertising. Venture capital firm Sequoia Capital believes AI businesses will eventually “trillions of dollarsat economic value.
In China, investors and entrepreneurs are racing to catch up.Stocks in Chinese artificial intelligence companies soaredThe flashy announcements were made by some of China’s biggest tech companies, including the recent e-commerce giant Alibaba; sense time, to create facial recognition software. and the search engine Baidu.At least two startups in development Chinese Alternative proposal OpenAI’s technology has raised millions of dollars.
ChatGPT is not available in China. But in the face of a growing number of homegrown alternatives, China has quickly revealed the artificial intelligence red line ahead of other countries still weighing how to regulate chatbots. .
The rule shows China’s “act fast and break things” approach to regulation, said Kendra Schaefer, technology policy director at Trivium China, a Beijing-based consulting firm.
“With no two-party system that both sides advocate, they can say ‘OK, we know we need to do this and we’ll fix it later,'” she added. .
Chatbots are being trained on a large scale of the internet, and developers sometimes grapple with inaccuracies and surprises in what they spit out. We are requesting a level of technical control over chatbots. Even companies like Microsoft are still fine-tuning their chatbots to eliminate harmful responses. The hurdles in China are much higher, so some chatbots have already been shut down and others are only available to a limited number of users.
the expert is Divided How difficult it is to train AI systems to be consistently fact-based. Some have wondered whether companies can account for the full scope of China’s censorship rules. Some even think that with enough work over time, machines can align themselves with truths and certain or even political values.
Analysts expect the rules to change after talks with Chinese tech companies. Regulators may relax enforcement of the rules so that they don’t completely undermine the development of the technology.
China has a long history of censoring the internet. Throughout the 2000s, the country built the world’s most powerful intelligence network on the web. It scared off Western companies like Google and Facebook. It has hired millions of workers to monitor Internet activity.
All the while, Chinese tech companies that had to abide by the rules thrived, defying Western critics who predicted that political domination would undermine growth and innovation. As technologies such as facial recognition and mobile phones emerged, companies helped nations use them to create a state of surveillance.
Matt Sheehan, an expert on AI in China and a fellow of the Carnegie Endowment for International Peace, says the current wave of AI poses new risks to the Communist Party.
The unpredictability of chatbots making nonsensical or false statements (which AI researchers call hallucinations) runs counter to the party’s obsession with controlling what they say online, he said. Sheehan said.
“Generative artificial intelligence has put two of the party’s most important goals: leadership in information management and artificial intelligence,” he added.
China’s new regulations aren’t entirely about politics, experts say. For example, it is intended to protect the privacy and intellectual property of individuals and creators of data used to train AI models. This is a global concern.
In February, image database company Getty Images sued artificial intelligence startup Stable Diffusion for training an image generation system on 12 million watermarked photos. Getty claimed it diluted the value of the image.
China is applying broader pressure to address legal issues related to the use of underlying data and content by AI companies. In March, Beijing established a National Data Authority as part of a major institutional reform. This is an effort to clarify the meaning of owning, buying and selling data. National agencies also help companies build the datasets needed to train such models.
“They are now deciding what kind of property data they have and who has the right to use and control it,” said Schaefer, who has written extensively on China’s AI regulations and has called the initiative “transformative.” It is called “Target”.
Still, China’s new guardrails may come at the wrong time. The country faces increasing competition and sanctions on semiconductors that threaten to undermine its competitiveness in technologies including artificial intelligence.
China’s hopes for AI ran high when AI engineer and entrepreneur Xu Liang unveiled one of China’s earliest answers in early February. Chat GPT as a mobile app. Xu said an app called ChatYuan was downloaded more than 10,000 times in his first hour.
Media reports of a marked difference between party policy and ChatYuan’s reaction quickly surfaced. The response offers a bleak diagnosis of China’s economy, describes Russia’s war in Ukraine as a “war of aggression” and is at odds with the party’s more pro-Russian stance. Days later, officials shut down the app. .
Xu said he is adding measures to create more “patriotic” bots. This includes filtering out sensitive keywords and hiring more manual reviewers to help flag problematic answers. He has trained another model that can detect the “wrong point of view” and filters it.
Still, it’s not clear when Xu’s bot will satisfy authorities. The app was originally set to reopen on February 13, according to the company. screenshotbut as of Friday it was still down.
It said “Services will be restarted once troubleshooting is complete”.
