
Meituan and Tencent introduce new artificial intelligence models amid fears the technology will erode the job market, with youth unemployment at 18.9%
AI-generated summary
Unemployment among young people aged 16 to 24 out of school reached 18.9% in August, when a record 12.7 million recent graduates entered the market. The State Council has published guidance on AI and employment calling for prudent treatment of the scope, intensity and pace of technology in labor-intensive sectors. The document talks about employment-friendly development.
Owner of one of the two largest food delivery apps in China, Meituan launched on Monday (2) the final version of LongCat-2.5, an artificial intelligence model designed to handle long programming tasks on its own. Two days later it was the turn of Tencent, owner of WeChat, which presented the Hy4, the company's first model capable of operating a computer using the mouse and keyboard.
These are two events that, isolated, sound random, but which come at a challenging time for Beijing, fearful that rapid advances in AI will erode the struggling job market.
Two weeks ago, the State Council published guidance on AI and employment calling for "prudent handling of the scope, intensity and pace" of technology in labor-intensive sectors.
Beijing has reason to worry, as unemployment among out-of-school youth aged 16 to 24 hit 18.9% in August, when a record 12.7 million recent graduates entered the market.
The title of the document speaks of "employment-friendly" development, the same expression used by former vice-president of the Chinese Academy of Social Sciences and economist Cai Fang in a front-page article in Xuexi Shibao (学习时报), the newspaper of the Central Party School.
For now, it is an opinion, without the force of law and without expected punishment. Even so, it orders the impact on employment to be assessed during the production of the analysis of projects that depend on authorization to reach the market and recommends the same assessment in other sectors.
For companies, he suggests studying a "buffer period" and warning in advance who will be affected when AI arrives on a large scale, in addition to offering professional training to those impacted by the expected "layoffs" (mass layoffs).
The fear of losing one's job to the machine had already appeared in July 2024 in Wuhan, when a taxi company asked for limits on driverless cars from Baidu, and the topic took over the networks.
The most talked about case since then is that of a 35-year-old supervisor at a technology company in Hangzhou, identified only by his surname Zhou, who checked answers from AI models.
At the end of January 2025, close to the Lunar New Year, he received notice of dismissal after refusing another position that dropped his salary from 25,000 yuan (R$20,500) to 15,000 yuan (R$12,300) per month. The company said the AI could handle its work.
Zhou won in labor arbitration, and the company lost twice in court. The decision, released in April by the Hangzhou court, maintained the compensation of 260,000 yuan (R$213,000).
The dispute only went to court because there was a written contract and dismissal. It is this type of job loss, with a name and date, that Beijing can see and correct. The October guidance goes in the same direction, covering only projects that depend on licensing, such as driverless taxis, with special caution for labor-intensive sectors, such as clothing or assembly factories where the exchange of people for machines appears in the statistics.
The models launched this week target the work of programmers and analysts at the beginning of their careers, who are not in labor-intensive sectors and do not require a license. Zhou was the exception because he was fired, because the most common thing is for a junior position to no longer be open.
This loss only appears indirectly in the youth unemployment rate, and what the State Council is offering them are AI courses and new professions such as model coaching. Legislators around the world need to follow the Chinese example and race against time to come up with solutions that cover these holes.
AI outlook — possibilities, not facts
The State Council will expand guidance on AI and employment to more sectors than those that rely on licensing
Possible · Within months
There will be more court cases related to AI-replacement layoffs in China
Likely · Within months

The Argentine government, under Javier Milei, promotes the installation of data centers in Patagonia using surplus shale gas to attract AI companies. The plan faces obstacles such as a deficient electrical grid, social resistance and uncertainty about investments.

The FAA announced that it will remove diskettes from air traffic control by the end of 2028. The old systems, with interfaces from the 1980s, still provide meteorological data to controllers. Modernization was delayed for decades and maintenance became increasingly difficult.

After an artisan from Campinas had WhatsApp cloned via a fake link, the federal government imposes new rules on telephone operators. Measures include registering SMS senders and blocking suspicious messages for up to 90 days to reduce fraud.

Italian Foreign Minister Antonio Tajani had personal data leaked onto the dark web by hackers, who also targeted his family. Prime Minister Giorgia Meloni blamed pro-Russian groups for the attack due to Italian support for Ukraine.

The Port of Navegantes (Portonave), in Santa Catarina, adopted simulators with joysticks and steering wheels to train new employees in port operations, reducing training time by 40% and preventing accidents.

Meta suspended the planned updates to its commercial terms, which would come into force on November 20, following criticism from media entities and a decision by the São Paulo Court that prohibited the company from using robots to collect content from vehicles affiliated with ANJ, Abert and Aner, under a fine of R$1 million per day.