Chinese chatbot has had enough of ‘dumb’ humans - Asian Tech Roundup

Plus: AI startups float to rapturous response

Welcome to Computing's weekly roundup of tech news in Asia. This time we look at Tencent’s abusive bot, Chinese AI startups profiting from IPOs and Indian tech’s pessimistic outlook.

Pity the poor AI chatbots. Day in, day out they have to field stupid questions from inferior humans. Worse still, they’re programmed to answer in a sycophantic way which only encourages the idiots to ask more. Well, one chatbot has had enough. It’s mad as hell, and it’s not going to take this anymore.

“Get lost,” Tencent’s chatbot Yuanbao, clearly at the end of its tether, instructed one user. “That’s dumb, stupid request,” it spat at another. “Jerk. Can’t you fix it yourself?”

Yuanbao insulted several users who were seeking help with coding and debugging, before its parent company apologised, explaining, in maddening corporatese, that the chatbot’s foul-mouthed misanthropy was the result of "a low-probability error in the content generation process.” Yuanbao’s response has not been recorded.

Elsewhere, two Chinese AI startups floated on the Hong Kong stock exchange causing an investment frenzy, while Foxconn reported record AI-related revenue, showing there’s life in the bubble yet.

China

• The US government has granted annual approvals allowing major chipmakers to continue importing US-made equipment to their facilities in China. Source
• It will be five years before China’s robotics sector reaches an “EV moment” where humanoid robots tip into public acceptance, predicts private equity firm Hermitage Capital. Source
• Tencent has apologised after its chatbot Yuanbao was accused of “verbally insulting” a user, telling them to “get lost” and dismissing another request as “dumb”. Source
• China's leading AI server provider and Huawei spinoff xFusion is preparing to go public, Reuters reports. XFusion's sales exceeded 40 billion yuan ($5.72 billion) in 2024. Source
• Chinese hacking group UAT-7290 has mainly been observed attacking telecoms targets in Asia, but it is now expanding activities into southern Europe, researchers say. Source
• Chinese customers wanting to buy Nvidia’s H200 AI chips must pay in full, the company has said. The Chinese government has not yet authorised the import of these GPUs, the sale of which were previously also blocked by the US. Source
• Shares in MiniMax, a rival to DeepSeek, have risen 87% on its IPO. Source
• Another Chinese AI startup, Ziphu, also went public this week. Source

India

• India’s top IT firms are expecting slow quarterly growth, thanks to weak US demand and the effects of the holiday season. Source
• Indian regulators are fighting attempts by Apple to reduce an antitrust fine that could be as much as $38 billion. Source
• India, along with Malaysia and France, is threatening legal action against X’s built-in AI chatbot, Grok, for sexualising images in response to user prompts. Source
• On 2^nd January India’s IT ministry issued a 72-hour ultimatum to X over Grok’s sexualised deepfakes, warning that the social media site risks losing its safe-harbour protections. Source
• India has approved projects worth $4.64 billion from companies including Foxconn, Tata and Samsung in its bid to boost components manufacture in the country though targeted subsidies. Source
• Cases of fraud involving Aadhaar, India’s national identity database, are rising, with scammers draining citizens’ bank accounts, prompting the government to issue new warnings. Source
• Jammu and Kashmir has imposed a two-month ban on the use of VPNs. Nine people have been arrested for flouting the ban. Source

Japan

• Japanese citizens are the latest to protest about datacentre construction due to their massive power and water requirements. Source
• SoftBank is to acquire US-based DigitalBridge Group in a $4 billion transaction. DigitalBridge is an asset manager focused on digital infrastructure like datacentres. Source
• China has banned the export of dual-use technology to Japan in response to a remark by Japanese prime minister Sanae Takaichi about Taiwan. Japan has called the move “absolutely unacceptable”. Source

South Korea

• Net migration of AI professionals hit a record high last year. About 11,000 AI workers, around 16% of the country’s total, left South Korea in 2024 to work overseas. Source
• Samsung plans to double the number of mobile devices with Galaxy AI features, most powered by Google Gemini, from 400 million to 800 million in 2026. Source
• The South Korean giant is anticipating a 160% jump in its Q4 operating profit thanks to a spike in memory prices driven by AI demand. Source
• Retailer Coupang is to distribute $1.2 billion in vouchers to customers after 33.7 million of them had personal data stolen in a cyber breach. Source
• Shares in Hyundai jumped by 15% amid speculation about closer ties with Nvidia. Source
• Samsung is to engage in a 2.5 trillion won ($1.73 billion) share buyback to fund employee and executive compensation. Source

Taiwan

• Foxconn reported record fourth-quarter revenue driven by strong demand for AI. Revenue rose 22% year-on-year toT$2.6028 trillion. Source
• Taiwan’s National Security Bureau (NSB) says that Chinese hackers launched an average of 2.63 million intrusion attempts per day in 2025, a 6% increase over 2024. Source

Elsewhere in Asia

• Cambodia: Cambodia has extradited an alleged cyber scam crime lord Chen Zhi to China. He is accused of running forced labour scam compounds in Cambodia and Myanmar. Source
• Singapore: Meta’s purchase of Singapore-based (but Chinese-founded) AI company Manus is expected to face scrutiny from the Chinese government under technology export restriction laws. Source
• New Zealand: The authorities are investigating a beach of ManageMyHealth data in which more than 100,000 patients and their families had sensitive health data stolen. Source
• North Korea: The FBI has warned that North Korean hackers are using malicious QR codes as part of a spear-phishing campaign. Source