Published: 17:56, June 7, 2024
New-age jobs
By Liu Yukun

Intelligent machines still need humans like data annotators, virtual architects to help them

(ZHANG YUJUN / FOR CHINA DAILY)

Completing the jigsaw of artificial intelligence (AI) requires a helping hand from humans, which is creating innovative professions and reshaping the employment landscape in the digital age.

As China’s AI industry continues its meteoric rise, it is not just technology that is evolving, but there are also new job opportunities that are blossoming in tandem.

Last year, the market size of the Chinese AI industry reached nearly 509.7 billion yuan ($70 billion), according to market research company iResearch. Soaring demand for jobs such as data annotators, virtual face designers and virtual architects highlights a crucial intersection where human skills complement machine capabilities.

The Ministry of Human Resources and Social Security received over 430 proposals last year for new occupations, more than double the number in 2021. About a quarter of those were closely tied to the digital economy, intelligent manufacturing, information technology and modern services. The newly recognized professions will be announced by the ministry in June.

“In the future, society will become increasingly virtual and digital,” said Wang Youwei, a professor of information management and business intelligence at Fudan University in Shanghai. “The industrial revolution freed people from repetitive physical labor, and generative AI is expected to solve repetitive mental labor. This will inevitably create new job opportunities.”

Wang elaborated on one example. “In content production-related jobs, such as document writing, and customer services that involve questions and answers, large models might eliminate some repetitive mental jobs. However, as these models can provide incorrect or illogical outputs, jobs involving judgment and monitoring of generated content will become more important.”

This “data annotation” was originally handled by software engineers and programmers. However, as data-cleaning demands grew, they became overwhelmed, giving rise to “digital gig workers” in the AI field.

Visitors are drawn to manufacturing equipment featuring AI technologies at an exhibition held in Suzhou, Jiangsu province, on June 2, 2024 (PHOTO / XINHUA)

Annotators identify specific words in text or speech, outline items in images or videos, and tag them. For instance, in synonym training, annotators determine if two text segments are semantically identical and mark “yes” or “no”. These tasks do not require high academic or professional qualifications, but do demand understanding, discernment and meticulousness.

Sun Danfeng, a stay-at-home mom from Zhoukou, Henan province, works as a data annotator. Despite having only a junior high school education and no knowledge of AI, she has become an AI trainer.

“There are four to five hundred data annotation projects on the platform I registered on. Some are simple, like selecting a different image from three pictures, and others are more difficult,” she said, adding that the job is flexible as annotators can choose what they do and when and where they work, as long as the task is completed on time.

“It’s very hard for a housewife who has no decent background in education to have a part-time job in a city like Zhoukou, or any non-mega cities,” she said.

“However, for stay-at-home moms, having nothing to do after the kids go to school is really boring, and asking for money from the husband is unpleasant.”

Sun learned about data annotation from her college student brother, when she spoke to him about feeling disconnected from society after becoming a full-time housewife. She lacked financial independence and struggled to find a job due to her limited education and the county’s tight job market.

“This data annotation job is perfect for me,” she said.

“Now, I can earn some pocket money in my spare time. I earn 1,000 to 2,000 yuan per month and am quite satisfied. My husband works in another city, and we have debts, so earning some extra money for the kids is good,” she added.

In 2020, the Ministry of Human Resources and Social Security included data annotators on the national occupational classification list, giving the job official recognition.

A visitor tries an AI information system at a digital fair held in Fuzhou, Fujian province, in April 2024. (PHOTO / XINHUA)

In recent years, the market size of data annotation has grown significantly, and is estimated to reach 20.43 billion yuan by 2029, up from 4.33 billion yuan in 2021, according to the Huaon Industrial Research Institute. This speedy growth has spurred ever greater demand for relevant skills.

However, Sun Ping, an associate researcher at the Chinese Academy of Social Sciences, expressed concerns about the rapidly expanding market.

“Data annotation companies, mainly outsourcing tasks and operating in a labor intensive mode with low entry barriers, pose a risk of easy replacement for their workers, earning them the moniker the ‘Internet’s Foxconn’. Annotators are both AI trainers and ‘job gravediggers’. When ultimate algorithms are developed, such a job may no longer be needed, leaving these workers once again unemployed,” she said.

Ye Cong, head of NetEase Yoo-Nexus Platform, has a more positive outlook on AI job creation.

“The digitalization of various industries will create continuous demand to train AI in different scenarios, offering more human-AI collaborative employment opportunities. The future relationship between humans and AI is one of collaboration, not replacement,” Ye said.

NetEase’s new platform accepts data tasks from businesses and breaks them down into multiple minor projects. Users can earn points by completing the projects, which can be exchanged for wages.

In less than three years since its foundation, the platform has garnered over 1 million registered users, mainly from western and northeastern regions of China, with an even gender split, mostly aged 25 to 35 years. Monthly active users exceeded 15,000.

The platform has expanded data annotation from traditional scenarios such as AI customer service, chatbots and image generators to physical projects in the engineering machinery industry, helping robots model tasks and providing more training scenarios and data for machine intelligence.

“We focus on excavators and loaders, helping them achieve precise smart operations in mines and ports,” Ye said.

“We’re now promoting collaboration with main engine manufacturers. The ‘AI-ization’ of physical industries ensures 24-hour excavator operation, greatly improving efficiency and addressing safety challenges in scenarios like underground mines,” he said.

In these projects, manual “annotation-based” guidance for machines to learn tasks places greater demands on data annotators. “Most of them need to have industry experience and many are certified excavator operators,” Ye said.

Jobs in the AI sector are on offer at a fair held in Beijing in May 2024. (WANG GUIBIN / FOR CHINA DAILY)

Beyond data annotators, new digital professions like face designers, motion capture artists, and virtual architects are also emerging. Instead of working at a company or platform, an increasing number of people are choosing to be self-employed.

College student Zhou Run is embarking on such a journey. He accepts orders on lifestyle platform Xiaohongshu to create virtual avatars that closely resemble real-life images.

For a fee of 50 yuan, clients can commission his work. However, to buy the rights to their avatar or obtain the numerical data of their avatar’s features, they need to pay around 600 yuan.

“Today more gamers prefer the ability to freely change their appearance and customize their characters to achieve a more realistic experience in the virtual world. The pre-designed faces in games no longer satisfy players’ needs for personalization, leading to the emergence of the ‘face sculptor’ profession, which is what I’m doing.”

Zhou has been in the business for just under six months and has already amassed several hundred followers on his Xiaohongshu account. During peak periods, he can earn 4,000 to 5,000 yuan a month.

As his order volume increases, he has transitioned from traditional modeling tools and face sculpting software to AI tools that allow for mass production.

VoiceAnimator is an AI-driven facial animation tool from NetEase Games AI Lab. This tool is capable of generating facial animations for virtual characters directly from inputted audio or text and audio files, that is suitable for hyper-realistic and anime-style virtual characters.

Such facial animations, which originally took 10 to 14 days to produce, can now be generated by AI in just two minutes. Based on the generated animations, artists can make further refinements.

“Creating vivid facial animations is indeed a complex, time-consuming, and costly endeavor. It requires understanding subtle facial muscle movements and variations, as well as smooth lip synchronization with speech. With the increasing prevalence of virtual characters, there’s a growing demand for efficient content production. We aim to facilitate this process,” said Yin Xu, the technical manager of VoiceAnimator.

“As the AI industry flourishes, continuous learning is essential for those within the industry. Professional artists are also embracing AI knowledge. VoiceAnimator users not only need professional artistic skills but also need to learn AI-related tools to reduce repetitive tasks and unleash their creativity,” Yin added.

A robot is presented at a digital expo held in Fuzhou, Fujian province, in May 2024. (PHOTO / XINHUA)

In the approximately four years since its development, this tool has been continuously optimized for various usage scenarios, from in-game character animations to educational virtual language tutors and AI-driven news broadcasting.

According to the Ministry of Industry and Information Technology, the number of AI companies in China has surpassed 4,500. It has spawned new industries and business models while creating countless job opportunities, including new roles for those in traditional sectors. Professionals who can adeptly combine AI technology with traditional industry knowledge are becoming highly sought after in the market.

Bella Zheng, a former landscape designer, is transitioning to become a “city architect” in the metaverse and immersing herself in learning coding.

“I want to create cities in the virtual world,” Zheng said. “Those who can do urban planning often can’t do metaverse, and vice versa. The new profession of metaverse city architect combines both skill sets.”

Many architects and space designers are joining the construction of the metaverse, freeing themselves from traditional constraints to creatively express unique works, according to Zheng. Renowned projects in the metaverse like Zaha Hadid Architects’ “Liberland” are bringing this new architectural horizon into view.

This profession aligns with the childhood dream of Zheng, who is currently based in Los Angeles, California, of traveling the world. “I used to do part-time projects, visiting different cities for inspiration, including domestic and international cities, understanding their iconic features and feeling the atmosphere, and we would integrate that into our projects,” she said.

What attracts her most to this job is the infinite creative space. “In the metaverse, rules or restrictions like gravity, structural stability, climate, or physical laws don’t apply. Architects can freely transcend existing conditions to create special environments and real artistic works,” she said.

“Years ago, parents might not have understood my profession and would have preferred that I become an urban designer, like most of my classmates did. But now their mindset has changed. This is an emerging career, which more and more people deem promising,” she said, adding that there is increasing recognition and acceptance, even from older people.

liuyukun@chinadaily.com.cn