China’s central internet regulator, the Cyberspace Administration of China (国家互联网信息办公室), issued proposed measures relating to the development and use of AI chatbots and other AI tools in China to solicit feedback from the public during the initial legislative drafting process. People can send in their views and comments on the draft until May 10 of this year.

The current draft, published on April 11 (translation here), explicitly focuses on managing the “healthy development” of generative artificial intelligence (AI) technology, referring to any type of AI system capable of generating text, images, or other media in response to prompts. Over the past few months, ChatGPT in particular has become a widely used tool across the world to generate human-like responses to a wide range of questions and topics, but services such as Midjourney have also become popular to generate images.

In China, ChatGPT has also triggered a lot of online discussions on the ways in which it could be used, the effect it will have on the labor market, and issues related to privacy, copyright, and censorship in China (read ‘ChatGPT in China‘).

Although users in mainland China officially cannot sign up to use Open AI’s ChatGPT, many people do find ways to use the platform. Earlier this year, while discussions about ChatGPT were prevalent in Weibo’s trending lists, Chinese tech giants announced their own plans to develop similar ChatGPT-like services: Baidu has Erniebot, Alibaba is rolling out Tongyi Qianwen, while Bytedance, Tencent, and Netease are also working on their own LLM (Large Language Models) chatbots.

As 2023 is already the year of the chatbot, it is perhaps unsurprising for China’s internet authorities to lay out the rules surrounding generative AI technologies.

There are a total of 21 sections or articles listed in the document. Many of the proposed rules are quite general and are about AI-generated content and ChatGPT-like services having to be in line with China’s overall internet and privacy laws.

The draft suggests that AI-generated content should “reflect the core values of socialism” (“应当体现社会主义核心价值观”), should not undermine the state authority nor the socialist system, cannot be harmful to national unity or social cohesion, and it also may not promote terrorism, extremism, discrimination, violence, obscenities, nor spread false information (article 1).

AI-generated content should not just be true and accurate, it should also ensure that the legitimate interests of others are respected and prevent harm to their physical and mental well-being, as well as damage to their reputation, privacy, and intellectual property rights (article 4).

In accordance with the Cybersecurity Law of the People’s Republic of China (中华人民共和国网络安全法), users of generative AI models should also register under their real name (article 9).

These kinds of AI service providers also should take clear and transparent measures to prevent users from becoming overly dependent or addicted to AI-generated content (article 10).

On Weibo, many comments relating to the proposed measures are supportive of them, as netizens especially express concerns over copyright issues and the problems surrounding ‘deepfake’ and AI services allowing users to swop faces or generate images using people’s faces.

But other people also think that when it comes to AI generated content, the rules are vague and hard to control – and comply with. How could AI-generated content always be “true and accurate” if the prompt given by a user, for example, is to create a fairytale or other fictional content? And how does one actually measure an “addiction” to AI services if they are part of a person’s everyday workflow? Some commenters fear that the rules could be arbitrarily applied because they are so broad and general.

Then there are those who think that AI services like the American OpenAI’s ChatGPT are developing so rapidly that China is already falling behind and that, especially in the light of these rules, ChatGPT will be much stronger than Chinese equivalents.

By having to embody socialist values and stay in line with strict Cyberspace rules, chatbot services will have to stay aligned with China’s traditional media and publishers. Some Weibo commenters discuss how ChatGPT also has bias and alleged anti-China sentiments. “We’ll have to add the ‘Party spirit’ to our AI,” one person replies.

“It’s the survival of the fittest,” another commenter wrote: “If there is no innovation, they’ll be eliminated.”

For a full translation of the “Measures on the Administration of Generative Artificial Intelligence Services (Draft for Solicitation of Comments)”, check out China Law Translate here.

By Manya Koetse , with contributions by Miranda Barnes

Follow @whatsonweibo

Get the story behind the hashtag. Subscribe to What’s on Weibo here to receive our newsletter and get access to our latest articles:

Subscribe

Spotted a mistake or want to add something? Please let us know in comments below or email us. First-time commenters, please be patient – we will have to manually approve your comment before it appears.

©2023 Whatsonweibo. All rights reserved. Do not reproduce our content without permission – you can contact us at info@whatsonweibo.com.

A work-related group chat that may or may not be real has gone completely viral on WeChat, Weibo, and beyond, triggering new discussions on China’s overtime work culture.

This week, an explosive WeChat conversation in which a worker confronted his team manager after being asked to work during the Qingming holiday attracted a lot of attention on Chinese social media.

The story started when screenshots of Chengdu’s China Electronics Technology Group Corporation (CETC/中国电子科技集团有限公司) work unit group chat allegedly leaked online and quickly spread on social media on the 4th of April. By nighttime, it had gone completely viral.

The group chat discussion was about a team manager who wanted his employees to work during the Tomb Sweeping (Qingming) Festival, which officially is a holiday.

“I need two people to work during Qingming, any volunteers? If not, I’ll appoint them,” the leader wrote. When nobody responded, the manager appointed a worker named Chen Zhilong (陈志龙) as one of the people who needed to work during the holiday.

Qingming, which was celebrated on Wednesday, is a Chinese festival when people traditionally visit ancestral tombs to sweep them and pay respects to ancestors and departed loved ones in other ways, such as by burning paper. It’s been an official public holiday since 2008.

Chen Zhilong then reacted angrily to the request in the group chat and refused to work overtime, saying staff members were already working long hours, some working from 8am to 11pm, and he asked the manager: “Did we sign a labor contract or a slave deed?” Chen, who called out the company out for illegal overtime and temporary work practices, resigned from his job and was supported by the other workers in the group chat.

As the story fermented online, CETC soon responded and denied that the group chat belonged to one of their work units and that the person named Chen Zhilong did not work for them. CETC also emphasized that the persons spreading this kind of fake news could be held accountable.

A related hashtag received over 230 million views on Weibo on Thursday (#中电科称痛批加班员工非集团公司员工#), while one Xinhua news post about the issue received one million (!) ‘likes.’

CETC denies the group chat is related to them.

So, are the screenshots fake or not? Some commenters think they have been photoshopped, while other sources – including Sichuan’s labor union – claim the incident actually happened last year already in a different city. Many commenters think the CETC is just trying to silence the topic.

Another document that supposedly was proof of Chen Zhilong working at CETC was officially debunked on Weibo.

Despite discussions over whether or not the group chat is real, many say it does not matter since the main issue is about the problem of being asked to do overtime work or work during holidays – which is a common issue within many Chinese companies.

“My company is rotten like this, they also made me fight for my free time during Qingming Festival to go and see my grandparents,” one commenter wrote, with another person replying: “Art comes from real life, and overtime work culture is very real. But virtually nobody has the courage to go against it.”

“It does not matter [if it’s real or not], I am still rooting for the guy scolding [the manager],” another popular comment said.

Over the past two to three years, there has been an increase in online discussions surrounding the strenuous work culture that is part of everyday life in many Chinese companies, especially tech and finance ones.

The topic has become especially newsworthy since 2021, when the deaths of two Pinduoduo employees sparked discussions on ‘996’ working culture. Later that year, Chinese authorities emphasized that the practice of ‘996’ (working 9am-9pm, six days per week) is illegal and that employers are obliged to obey the national working-time regime.

Although these kinds of overtime practices are technically prohibited by law, many companies still enforce the hours informally and in online discussions, many workers revealed that they were still assigned job tasks that exceeded the prescribed working hours.

The Wechat work group conversation atrracted so much attention because it resonated with many netizens, and also because it was suggested to be about CETC, a Chinese state-owned company.

Read more about the ‘996’ working culture here.

By Manya Koetse, with contributions by Miranda Barnes

Follow @whatsonweibo

Get the story behind the hashtag. Subscribe to What’s on Weibo here to receive our newsletter and get access to our latest articles:

Subscribe

Spotted a mistake or want to add something? Please let us know in comments below or email us. First-time commenters, please be patient – we will have to manually approve your comment before it appears.

©2023 Whatsonweibo. All rights reserved. Do not reproduce our content without permission – you can contact us at info@whatsonweibo.com.