AI could automate as many as 300 million full-time jobs

As professions around the world assess the near-future implications of emerging generative artificial intelligence (AI), in China, accountants have begun calculating the ways that ChatGPT-like products could be incorporated into their field.

Analysing a company’s business status, identifying financial risks and preparing financial reports are some examples of how generative AI services could take on some of the workload of the professionals, wrote Liu Qin, a professor at the Shanghai National Accounting Institute, in an article in the journal Friends of Accounting.

“Accountants should embrace new technologies and boost their work efficiency with AI systems while cultivating creativity,” Liu said. “This is how they can hold an invincible position in the workplace.”

While some professions are enthusiastic about how using AI will enhance job efficiency, some roles could be put at risk as AI develops. When asked by the Post, ChatGPT responded that the key will be to adapt to the changing work environment with new skills.

A study published in March found that 54 per cent of jobs in China are at high risk of being replaced by AI in the coming decades.

“Relatively speaking, unit heads are the safest jobs in China, whereas jobs intensive in perceptive and manipulative tasks are highly susceptible to substitution,” researchers at the Central University of Finance and Economics in Beijing and the University of Hull in Britain wrote in the peer-reviewed journal Finance Research Letters.

But the team said workers in product manufacturing and business services are more likely to be hit, according to the researchers.

“As jobs evolve alongside the development of AI, workers will need to adapt to activities that require more creativity and social intelligence,” they said.

“Policymakers will need to provide more mid-career job training and income support for workers caught in the flood of AI to fulfil job transition.”

Julia Chen, an associate at Concordia, a Beijing-based consulting firm focused on supporting the safe and beneficial development of AI, said generative AI is set to play a significant role in the future of work in the world’s two largest economies.

“An important question that will affect adoption is how far and how quickly challenges associated with large language models, such as hallucination and bias, can be addressed,” Chen said.

She said another area to keep an eye on is to what extent government intervention will moderate the impacts on the job market.

“Chinese regulators have proposed strict regulations for generative AI models, which contain some sensible measures, but may slow investment and deployment in their current form,” she said.

“Calls for regulation in the United States from experts and the public are increasing, but it seems challenging to build political consensus around what that should look like.”

Earlier in April, China’s internet watchdog rolled out draft rules for ChatGPT-like services, which require developers to prevent discriminatory content, false information and content that infringes on personal privacy or intellectual property. All generative AI products must also pass a government security assessment before being released to the public.

Meanwhile, the US government is seeking public comment on how to create accountability measures for AI to ensure privacy and transparency.

Since Microsoft-backed OpenAI launched ChatGPT, big Chinese tech firms have been rushing to push out competing technology for the domestic market. They include Baidu, Tencent, SenseTime, iFlyTek, Huawei and Alibaba – the owner of the South China Morning Post.

While the American product is not officially available in China, internet users have found ways to access it and many sectors have expressed interest in using similar technologies.

As both countries compete to lead in the AI race, Chen said she is worried that fear of the other “catching up” or “racing ahead” could be exploited to accelerate investment and progress in generative AI systems without due consideration for good governance.

“Instead, what I’d really like to see is a race to the top, in terms of how to intelligently govern generative AI. It’s in the common interests of both countries to harness the productivity benefits of these systems while managing the risks of societal and economic disruption,” she said.

In a paper posted ahead of peer review, researchers at OpenAI found that jobs that rely heavily on science and critical thinking skills were less exposed to risks of automation, as opposed to programming and writing skills.

The team found that around 80 per cent of the US workforce could see at least 10 per cent of its tasks affected.

It listed jobs in which AI language models could save workers “a significant amount of time completing a large share of their tasks”, including mathematicians, accountants, journalists, clerks, blockchain engineers, climate change policy analysts and clinical data managers.

“With access to a LLM [large language model], about 15 per cent of all worker tasks in the US could be completed significantly faster at the same level of quality,” the team wrote, adding that that share jumps to 56 per cent of all tasks when software and tools built on top of LLMs are used.

In a global analysis published in late March, Goldman Sachs estimated that generative AI could automate as many as 300 million full-time jobs.

The US investment bank found that one-quarter of current tasks could be automated by AI in the US and Europe, listing sectors such as office administrative support, legal, architecture and engineering.

“The good news is that worker displacement from automation has historically been offset by creation of new jobs, and the emergence of new occupations following technological innovations accounts for the vast majority of long-run employment growth,” the firm said.

In China, a low birth rate combined with a rapidly ageing population is resulting in a widening demographic imbalance.

While generative AI could automate certain tasks and replace humans in some jobs, it might also help ease pressures produced by a shrinking young workforce in the country.

At the China Development Forum last month, a high-profile event attended by Chinese leaders and dozens of foreign CEOs, Zhou Hongyi, founder of Chinese cybersecurity firm 360 Security Technology, predicted that a new profession of AI trainers will emerge in China.

“Only with good teachers can AI be taught to hear one word and understand two,” he said, adding that his company had begun inviting users to register for an internal test of its ChatGPT rival “360 Smart Brain”.

“Young people will not become obsolete. It is those who still have not used GPT, do not know how to use GPT or do not know how to ask questions to GPT who will fall behind.”

In the US, an AI teacher role has already appeared. Anthropic, an American firm backed by Google parent company Alphabet to create AI systems, has been looking to hire a “prompt engineer”.

“You will figure out the best methods of prompting our AI to accomplish a wide range of tasks, then document these methods to build up a library of tools and a set of tutorials that allows others to learn prompt engineering,” states a job advertisement for the position that pays up to US$375,000 annually.

When ChatGPT was asked by the Post if it will replace humans in jobs, it replied: “While AI may replace certain tasks within jobs that are repetitive or can be easily automated, it is unlikely to replace humans entirely across all job sectors.

“Instead, AI has the potential to enhance and complement human work, allowing people to focus on more creative, complex and strategic tasks.”

It also offered career advice to humans.
“In many cases, AI will lead to job transformation, requiring people to develop new skills and adapt to the changing work environment.

“The key to minimising job displacement is to focus on reskilling and upskilling workers, preparing them for the jobs of the future that will be shaped by AI and other emerging technologies.”sc

end item*

AI could automate as many as 300 million full-time jobs

Differentiating the effect of social enterprise activities on health

Making norms to tackle global challenges: The role of Intergovernmental Organisations

American Militarism, Its Agents, And Its Next Target: Iran