-
TOPIC

ChatGPT and other generative AIs

ChatGPT and other generative AIs
Latest news and updates on ChatGPT, a large language model conversational bot operated by Microsoft-backed OpenAI that is able to understand sophisticated questions and give surprisingly humanlike text responses, and other generative AI, including Baidu's Ernie Bot, and image generation software such as Midjourney, Stable Diffusion and Dall-e 2. The articles focus on the implications of such technology, the challenges it could pose China’s censors, possible bans by educational institutions, AI regulation, and the development of other similar software.
T-Head, the semiconductor design arm of Alibaba, promotes its Zhenwu 810E chip as capable for both AI training and inference. Photo: Handout
Semiconductors

Alibaba’s T-Head unit unveils details of AI chip designed to rival Nvidia’s GPUs

The launch of the Zhenwu 810E marks the latest initiative by China’s Big Tech firms to develop a domestic alternative to Nvidia’s AI chips.

Artificial intelligence

DeepSeek taps Alibaba open-source AI technology to boost OCR performance

Artificial intelligence

China’s Alibaba, Moonshot claim latest AI models challenge OpenAI, Google

Education in Hong Kong

AI mental health app helps Hong Kong students accentuate the positive

EmoCare, powered by artificial intelligence, began trial run last month, serving around 700 primary, secondary and university students.

EdUHK’s AI Education Tools showcase with (left to right) Professor Alpha Ling Man-ho, Professor Song Yanjie, and associate vice-president (quality assurance) of the department of psychology Susanna Yeung Siu-sze. Photo: Jonathan Wong
Artificial intelligence

How China’s AI video generator Kling challenges Google’s Veo, OpenAI’s Sora

With Kuaishou’s tool headed for US$240 million in annual revenue, the firm is one of the world’s most undervalued AI stocks, JPMorgan says.

European Union

EU opens probe into Musk’s Grok over sexual AI deepfakes

Brussels targets X with a new probe into ‘digital undressing’, vowing to protect women and children from illegal AI content.

Advertisement
Advertisement
Advertisement
South Korea

Google’s Gemini surges in South Korea, challenging ChatGPT, after latest launch

South Korea generated about US$21 million, or 11.4 per cent of Gemini’s cumulative revenue on Apple’s app store – second only to the US.

A Gemini 3 logo is displayed on a smartphone screen on November 19, 2025. Photo: CFOTO/Future Publishing/Getty Images
Artificial intelligence

Open Questions | Jeffrey Ding on why diffusion, not innovation, is the secret to AI success

Professor says a ‘diffusion marathon’, where AI is gradually embedded in an economy over decades, will determine the winner of US-China tech war.

videocam
South Korea

South Korea begins investigation into Grok, sexually exploitative images

Grok is estimated to have generated around 3 million sexually explicit images between December 29, 2025, and January 8 this year.

Chinese culture

What does AI predict will happen in the fiery Year of the Horse 2026?

With the rise of AI mirroring the Year of the Fire Horse’s restless energy, we ask DeepSeek, Perplexity and Gemini their 2026 predictions.

Alphabet Inc.’s Google debuted an updated version of its artificial intelligence model, Gemini, that executives said represents a “massive jump” in reasoning and coding ability. (Dreamstime/TNS)
Artificial intelligence

Computing constraints force China’s Zhipu to restrict AI coding service sign-ups

The issue reflects the challenge faced by Chinese AI firms to expand their businesses amid computing infrastructure constraints.

Zhipu AI’s GLM Coding Plan service was launched in September. Photo: Shutterstock
Artificial intelligence

Baidu launches Ernie 5.0 as firm’s AI assistant users reach 200 million a month

The Beijing-based company’s most advanced AI model can process text, images, audio and video.

Artificial intelligence

Open Questions | Why practical use of AI, not race for supremacy, is key for this author

Academic Jeffrey Ding explains why human capital will determine whether US or China wins the AI race, not which attains superintelligence first.

videocam
Education in Hong Kong

Hong Kong teams up with experts to develop AI tailored for schools

Secretary for Education Christine Choi tells legislature authorities are exploring launching a large language model relating to city’s school curriculum.

5
Students use an AI platform powered by Microsoft for their study. Photo: Kelly Le
Asean

As more nations block Grok for obscene content, calls for Asean-wide curbs mount

The Philippines should push the fight against deepfakes high on the Asean agenda as the bloc’s chair, analysts say.

The Grok AI website displayed on a mobile phone as a nationwide restriction on the service took effect in Malaysia earlier this month. Photo: EPA
DeepSeek

DeepSeek a year on: how a Chinese start-up sparked a global AI arms race

DeepSeek’s meteoric rise propelled the growth of the open-source AI ecosystem by spurring its competitors to join the movement.

videocam
Artificial intelligence

Tencent seeks cooperation with other AI firms to better serve vulnerable users

Tencent Research Institute looks to build expert data sets for AI models that could broadly benefit the world’s second most populous nation.

Advertisement
Advertisement
Advertisement
X (formerly Twitter)

Musk’s xAI sued over Grok deepfakes – this time, by mother of his child

Ashley St Clair, mother of Musk’s 13th child, said that in one photo, an image of her at 14 was altered to show her in a bikini.

X owner Elon Musk and Ashley St Clair. Photo: X/stclairashley
Malaysia

X marks the bot: Malaysia says Grok ban may be lifted if it is proven to be safe

Though X has deactivated several prompts, more changes need to be made to ensure the site’s safety, Malaysia’s communications minister says.

Artificial intelligence

Musk’s Grok barred from undressing AI images after global backlash

‘Spicy Mode’ feature had allowed users to create sexualised deepfakes of women and children.

Artificial intelligence

Alibaba reaffirms open-source AI commitment as it hails successes

‘Building with the community’ is crucial to the success of the Qwen family of large language models, company says.

Leaning into the open-source approach has helped developers build AI applications faster and at lower cost, company says. Photo: Shutterstock
Education in Hong Kong

Nearly 1 in 4 Hong Kong students can’t finish homework without AI, study shows

Our Hong Kong Foundation calls for unified AI platform and more comprehensive curriculum to cover students at all academic levels.

7
AI apps on a smartphone screen: ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot, Meta AI, Grok and DeepSeek. Photo: dpa
Hong Kong economy

Producers at Hong Kong toy fair bet on AI-powered plushies to double sales

But some buyers also express concerns over use of the technology in toys for children.

Artificial intelligence

Ant International teams up with Google on use of Universal Commerce Protocol

The endorsement of Google’s open standard comes as companies globally accelerate the integration of AI agents and e-commerce.

videocam
Artificial intelligence

Alibaba’s Qwen family hits 700 million downloads to lead open-source AI adoption

Alibaba Cloud’s Qwen family is the world’s most widely used open-source AI system as of January, Hugging Face data shows.

The volume of Qwen downloads on Hugging Face reinforced its position as the world’s leading open-source ecosystem. Photo: Shutterstock
Artificial intelligence

China’s AI unlikely to surpass US soon despite progress, researchers say

China’s top AI scientists warn of chip deficits and resource limits, though some predict breakthroughs within five years.

14
A humanoid robot is seen at the 2025 Global Industrial Internet Conference in Shenyang, Liaoning province. Photo: Xinhua
Elon Musk

Elon Musk to open source X algorithm in 7 days amid global scrutiny

The company has been working to incorporate more AI into its recommendation algorithm through Grok, Musk’s AI chatbot.

Artificial intelligence

DeepSeek to customise namesake chatbot for Italian users following probe

The firm agrees to a package of commitments, including tackling AI model ‘hallucination’ issues, with the Italian Competition Authority.

Advertisement
Advertisement
Advertisement
Help preserve 120 years of quality journalism.
SUPPORT NOW
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement
Advertisement