Beijing to Release Manus, DeepSeek’s Next Generation

The DeepSeek logo is seen in this illustration taken on January 29, 2025. (Reuters)
The DeepSeek logo is seen in this illustration taken on January 29, 2025. (Reuters)
TT
20

Beijing to Release Manus, DeepSeek’s Next Generation

The DeepSeek logo is seen in this illustration taken on January 29, 2025. (Reuters)
The DeepSeek logo is seen in this illustration taken on January 29, 2025. (Reuters)

Chinese artificial intelligence startup Manus on Thursday registered its China-facing AI assistant and was featured for the first time in a state media broadcast, highlighting Beijing's strategy of boosting domestic AI firms that have received overseas recognition.

Since China's DeepSeek shocked Silicon Valley by releasing AI models comparable to its US competitors but developed at a fraction of the cost, Chinese investors have been on the lookout for the next domestic startup with the potential to upend the global tech order.

Some have pointed to Manus. The company went viral on X a few weeks ago by releasing what it claimed to be the world's first general AI agent, capable of making decisions and executing tasks autonomously, with much less prompting required compared to AI chatbots like ChatGPT and DeepSeek.

Beijing is now showing signs that it will support Manus' rollout within China, echoing its response to DeepSeek’s success. State broadcaster CCTV on Thursday devoted television coverage to Manus for the first time, publishing a video on the difference between its AI agent and DeepSeek's AI chatbot.

Beijing's municipal government on Thursday announced that a Chinese version of an earlier Manus product, an AI assistant called Monica, had completed the registration required for generative AI apps in China, clearing an important regulatory hurdle.

Chinese regulators require all generative AI applications released in the country to abide by strict rules, partly designed to ensure these products do not generate content considered sensitive or damaging by Beijing.

Last week Manus announced a strategic partnership with the team behind tech giant Alibaba's Qwen AI models.

The move could bolster the domestic roll-out of Manus' AI agent, which is currently only available to users with invite codes and has a waiting list of 2 million, according to the startup.

In the markets, Hong Kong and China stocks declined on Friday and registered weekly losses, as tech shares tumbled on mounting profit-taking pressure.

The Hang Seng Tech Index slid 3.4% on Friday, and Hong Kong's benchmark Hang Seng Index lost 2.1%. Both indexes registered back-to-back weekly losses for the first time since January.

In Hong Kong, chipmaker Semiconductor Manufacturing International Corporation slid 7.5% to a one-month low, while market heavyweight Alibaba lost 3.5%.

China's blue-chip CSI300 index dipped 1.5%, ending the week with a 2.3% loss in its largest retreat since January. The Shanghai Composite Index lost 1.3%.

The tech sector also paced declines onshore. Mainland's tech-focused Star 50 Index dropped 2.1% and AI-related shares slipped 3%.

"It's normal to see some pullbacks at these levels after such a strong rally this year - this doesn't even qualify as a correction," said Dickie Wong, Kingston Securities executive director.

The optimism around China's “two sessions,” DeepSeek and President Xi Jinping's meeting with tech leaders has already been priced in with major indexes at current levels, prompting investors to take profit, he added.

The Hang Seng Tech index has lost 4.1% this week in a second week of decline - the longest losing streak since the opening weeks of the year.

However, the gauge is still up 26% year-to-date.



Copyright Questions Loom as ChatGPT's Ghibli-Style Images Go Viral 

Hayao Miyazaki of Japan, director of the animated film "Ponyo," poses at a special screening of the film in Los Angeles on July 27, 2009. (AP)
Hayao Miyazaki of Japan, director of the animated film "Ponyo," poses at a special screening of the film in Los Angeles on July 27, 2009. (AP)
TT
20

Copyright Questions Loom as ChatGPT's Ghibli-Style Images Go Viral 

Hayao Miyazaki of Japan, director of the animated film "Ponyo," poses at a special screening of the film in Los Angeles on July 27, 2009. (AP)
Hayao Miyazaki of Japan, director of the animated film "Ponyo," poses at a special screening of the film in Los Angeles on July 27, 2009. (AP)

The release of the latest image generator on OpenAI's ChatGPT has triggered an online flood of memes featuring images done in the style of Studio Ghibli, the Japanese studio behind classic animated films like "My Neighbor Totoro" and "Princess Mononoke."

The virality of these images, with OpenAI CEO Sam Altman even changing his profile picture on X to match the style, immediately raised questions about copyright infringement by the ChatGPT maker, which already faces lawsuits regarding the use of source material without permission.

Since the release on Wednesday, AI-generated images depicting Studio Ghibli versions of Elon Musk with US President Donald Trump, "The Lord of the Rings," and even a recreation of the September 11 attacks have gone viral across online platforms.

On Thursday, the White House took part by posting on X a Ghibli-style image of a weeping alleged felon being handcuffed by a US immigration officer before her deportation.

Originally intended to be available on the platform for free, Altman said the huge success of the new generator was unexpected and meant the tool would remain limited to paid users for now.

It was already possible to generate images with ChatGPT, but the latest version is powered by GPT-4o, the company's highest-performing model, and allows sophisticated results to be obtained through very succinct requests, which was not the case before.

After the viral trend, a video from 2016 resurfaced in which Studio Ghibli's legendary director Hayao Miyazaki is seen lashing out during an AI demonstration by staff.

"I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself," an English translation of his remarks said in the video.

The trend "is especially insidious and malicious because of how outspokenly scathing Miyazaki has been toward the tech," wrote artist and illustrator Jayd "Chira" Ait-Kaci on Bluesky.

"It's always about contempt for artists, every time," Ait-Kaci added.

OpenAI is facing a barrage of lawsuits over copyright infringements, including one major case with the New York Times and others from artists, musicians and publishers.

The company is aggressively lobbying the White House and Congress to make the use of copyrighted content by AI companies part of the fair use doctrine.

Fair use allowances already apply to search engines or in the case of satire and memes online, and allow companies to freely use copyrighted material without permission.

Bloomberg reported on Wednesday that OpenAI is close to finalizing a $40 billion funding round led by Japan's SoftBank Group that would be the biggest funding round ever for a startup.

OpenAI has projected its annual revenue could exceed $12.7 billion in 2025, up from $3.7 billion expected this year.