Redian新闻
>
A ChatGPT Gag Gone Wrong, a Police Probe, and a Sheepish Apology

A ChatGPT Gag Gone Wrong, a Police Probe, and a Sheepish Apology

社会

A ‘report’ on Hangzhou canceling traffic restrictions went viral. Then it emerged that a local resident used the AI chatbot to write it.

What started as a joke about using ChatGPT to mimic a government announcement canceling traffic restrictions ended in massive public confusion and a police investigation. 

It began when police in Hangzhou, in the eastern Zhejiang province, were forced to respond to widespread rumors that stemmed from a screenshot of a “notice” that emerged online Thursday. It stated that on March 1 the city would end a road space rationing policy, which restricts vehicles from driving on roads based on their license plate number. 

The “notice” not only listed three detailed reasons but also said: “Though such a policy reduced congestion to a certain extent after years of implementation since 2006, it has also caused public inconvenience.” 

Incidentally, the city suspended its traffic restrictions last December with the exception of expressways and said that restrictions would resume in February depending on traffic conditions at the time.

But on social media, the fake announcement went viral. Only later did local media outlets realize that the notice was written by ChatGPT, which has become an internet sensation in China since its launch. 

The state-run Zhejiang People’s Radio Station claimed Friday that the screenshot originated from a residential compound. There, a resident asked ChatGPT to write a “news report” about the cancellation of the road restriction policy, documented the entire process, and shared screenshots of the AI chatbot’s responses to a group on WeChat, China’s social superapp. 

“The city government hasn’t announced changes to the policy at all,” the report stated, adding that the resident had since apologized to the group where he shared the “screenshot.”  

The incident triggered an uproar on social media, as many expressed concerns over the chatbot’s capabilities. “It is really scary that it was literally written by AI,” stated a highly upvoted comment on Weibo. 

With its convincing ability to generate human-like responses, the popularity of ChatGPT has quickly surged across China in the past weeks, despite not technically being available for use in the country. Nevertheless, many have found a way of accessing the service, with some even seeking to use the chatbot to their own advantage.

But the advanced technology and its growing applications have drawn concerns from legal experts in China, who warned that, since it generates content based on massive data sets, it could become a powerful source of rumors and misinformation in the absence of a fact-checking mechanism. They worried that the lack of regulations would make it difficult to hold the offenders accountable.

“The content generated by AI technologies is not always correct and it needs some professional knowledge to figure out whether the information provided by ChatGPT is true or not,” Xie Lianjie, a Beijing-based lawyer specializing in internet law, told domestic media. He added that such tools could be trained by their developers to generate fraudulent information or provide phishing websites.

China’s internet regulators have vowed to increase their vigilance of online rumors and misinformation by putting the emphasis on rooting out their origins. In 2022, the Cyberspace Administration of China handled 172 million cases involving the spread of illegal and inappropriate information, 3.6% higher than the previous year.

Editor: Apurva 
(Header image: VCG)


Download the new Sixth Tone app at the App Store or Google Play
APK file for Android:
https://image4.sixthtone.com/pkg/sixthtone.apk
(Copy URL and open in browser)

微信扫码关注该文公众号作者

戳这里提交新闻线索和高质量文章给我们。
相关阅读
停止内耗!Google和DeepMind被ChatGPT逼急了,决定合作打败GPT-4Prompt Engineering全面自动化:LeCun看了沉默,ChatGPT看了直呼内行青拌柠檬大虾!Postdoctoral Research Associate in Immunology and Infectious Dis[电脑] The Grand Beyond The Grand —— 华硕ROG HYPERION创世神 装机SHOW!今天唱了《那些花儿集成ChatGPT后必应日活量首破亿!微软推出Azure OpenAI ChatGPT 服务,GPT-4下周发布狮王的一生,终有一战 (塞伦盖提最漂亮的狮子王 gone)Lee Wong | Sherry Chen Event Talking Points今天的三餐。C罗和梅西对饮料的节制JAMA Neurology: 独创介入式脑机接口,瘫痪患者的福音?剽窃、作弊,ChatGPT竟偷偷生成了12篇署名论文!Marcus炮轰Ta变成CheatGPTKinder Gone: As Enrolments Dip, Preschools Fight to Stay Afloat圣地亚哥散记ClickPrompt:一站式 Prompt 学习、设计与运行工具(开源),支持 ChatGPT 等斯坦福5.5万人研究:30年,男人长(cháng)了3厘米,但并非好事一文详解Prompt学习和微调(Prompt Learning & Prompt Tuning)chatGPT. 一本正经胡说八道ChatGPT讲座:ChatGPT并不神秘,但ChatGPT很神奇Hepatology | 李于研究组发现调控星状细胞激活与非酒精性脂肪性肝炎的新机制首次!TikTok CEO周受资将在美国国会发言;SHEIN再降高管,为软银前COO;拼多多奔向加拿大丨Going GlobalThe Answer to China’s Energy Woes? Harnessing Virtual Power.Chatgpt 4 要release 了反击!Google 版 ChatGPT 首次亮相,有一个功能「碾压」ChatGPTCan Archaeology Prove China’s Ancient Historians Right?Chinese Professor Suspended For Racist, Homophobic Speech内资会计所 | RSM 2023 Technology Risk Consulting Intern,持续热招中Prompt Engineer也保不住工作了!Midjourney发布图片生成Prompt功能“Describe”巴黎,巴黎(7)[录取捷报]港中文 Information Science and Technology Management 专业录取 来啦!After Losing Her Son to a Chicago Shooting, She Tries to Move OnGuangzhou Postpartum Care Center Probed Over Surrogacy Claims人手一个ChatGPT!微软DeepSpeed Chat震撼发布,一键RLHF训练千亿级大模型从病毒清零烂尾看大陆社会,体制和道德的沉沦Microchip Technology:半导体行业的希望之星PK chatgpt 的诗评读名言学英语:Writing is thinking. To write well is to think clearly.微软宣布开源DeepSpeed-Chat ,训练提速15倍以上,有望实现人手一个ChatGPT?微软开源DeepSpeed Chat,人人可快速训练百亿、千亿级ChatGPT大模型The Story of China’s Largest Genealogy Collection
logo
联系我们隐私协议©2024 redian.news
Redian新闻
Redian.news刊载任何文章,不代表同意其说法或描述,仅为提供更多信息,也不构成任何建议。文章信息的合法性及真实性由其作者负责,与Redian.news及其运营公司无关。欢迎投稿,如发现稿件侵权,或作者不愿在本网发表文章,请版权拥有者通知本网处理。