HONG KONG — Apple has removed more than a hundred apps offering ChatGPT-like services from its China store, as a regulation on generative artificial intelligence (AI) is set to take effect in the country in two weeks.

The apps were all pulled off the shelves from China’s iOS app store on Tuesday (Aug 1), according to data on Chinese mobile app analytics platform Qimai.

Spark, an app developed by iFlyTek that provides ChatGPT-style services, was among the apps taken down even though it had a high-profile launch on June 29.

Spark reappeared in China’s iOS app store on Wednesday morning, after iFlytek said on an investor relations platform that the app had gone through an update.

Among the other apps pulled was the popular ChatGAi Plus — which provides chatbot, AI translation and writing services — that was ranked 9th on the China iOS app store’s paid app chart before being removed on Tuesday afternoon, Qimai records show.

In a notification to developers, Apple said that it had removed the apps “because they include content that is illegal in China”, citing tighter regulations on deep synthesis technologies and generative AI.

This is according to a screenshot shared on X — formerly Twitter — by developer Zhenlu Zuo, who was behind one of the removed AI apps OpenCat.

In the screenshot, Apple also urged developers to comply with The Administrative Provisions on Deep Synthesis for Internet Information Service which came into effect in January — including securing a licence from China’s Ministry of Industry and Information Technology (MIIT), which allows them to offer such services. 

Apple did not immediately respond to a request for comment.

The move comes as enthusiasm for generative AI remains high in China while Beijing seeks to ramp up scrutiny of the sector.

Last month, China issued the world’s earliest and most detailed regulation on generative AI models, drawn up jointly by seven Chinese regulators including the Cyberspace Administration of China (CAC) and the MIIT.

The regulation, slated to take effect on Aug 15, targets all generative AI content services, including text, pictures, audio and video. It requires firms that are hoping to offer their products to the general public to promote healthy content and “adhere to core socialist values”, and to not generate false information or any content that endangers national security.

According to the new rules, many providers will need to pass a security assessment with the CAC before serving the public. It also requires companies to avoid any discrimination based upon ethnicity, faith, country, region, gender, age, occupation and health, in terms of the design of algorithms and the selection of training data.

Chinese companies are racing to produce their own answers to Microsoft-backed OpenAI’s ChatGPT amid an investment frenzy.

During the first half of this year, 22 generative AI startups in China received funding, the largest number in the world, AI-focused research firm Zhidongxi said in a report last month.

The United States was close behind with 21, though American startups are better funded than their Chinese peers.

But so far no domestic firm has rolled out a major, public facing ChatGPT-style service in China.

Some of the most high-profile endeavours by China’s Big Tech firms, including Baidu’s Earnie Bot and Alibaba Group Holding’s Tongyi Qianwen, are either still in trial mode or targeting enterprises only. Ernie Bot is still on Chinese app stores and can be downloaded by any user who has an invitation code to use it.

Alibaba owns the South China Morning Post. SOUTH CHINA MORNING POST