Machine learning: jobs that may be at risk as AI tools grow in importance

Scottish Revival singer-songwriter Ewan McCall’s 1986 song “My Old Man” was an ode to his father, a steel worker whose job was threatened by the advent of technology. The lyrics were written by industry leaders and tech stalwarts. GPT for OpenAI-4 and the ability to write essays, code, and perform calculations will be more accurate and consistent, heralding a fundamental technological shift. Almost as important as creating integrated circuits, personal computers, web browsers, or smartphones. However, the impact of advanced chatbots on the job market remains a question mark. If blue-collar jobs were the focus of his MacColl Ballad, generative pre-trained transducer-type artificial intelligence (AI) models represent a greater threat to white-collar workers. Operations on arrays of inputs end up producing very human-like output. So will this latest wave affect current employment levels?

trade off

According to Goldman Sachs economists Joseph Briggs and Debesh Kodnani, the answer is an overwhelming yes, with 300 million full-time jobs worldwide being “automated” and workers being replaced by machines and AI systems. I expect it to be replaced. Giving credence to this grim prediction is the new wave of AI, especially large-scale language models that include neural networks such as his Microsoft-backed OPenAI his ChatGPT.

Economists at Goldman Sachs have predicted that such technology could create “significant disruption” to the labor market, with lawyers, economists, writers and administrative staff most at risk of being laid off. predicted to be high. In a new report, “Artificial Intelligence’s Potentially Big Impact on Economic Growth,” they found that about two-thirds of jobs in the US and Europe are set to be “exposed” to varying degrees of AI automation. I’m calculating.

White-collar workers in general, and workers in developed countries in general, are projected to be at greater risk than blue-collar workers in developing countries. “The combination of significant labor cost reductions, new job creation, and increased productivity of the unemployed workers will lead to labor productivity levels that have followed the advent of early commodity technologies such as electric motors and personal computers. It increases the chances of a boom,” said the report.

And OpenAI itself predicts that the majority of workers will have at least some of their jobs automated by the GPT model. In a study published on the ‘arXiv’ preprint server, OpenAI and University of Pennsylvania researchers found that 80% of the U.S. workforce could be ‘affected’ by the introduction of GPT, with at least 10% of tasks said.

Central to these predictions is how models such as ChatGPT will improve with more use. GPT stands for Generative Pre-trained Transformer and is a marker for how the platform works. Initially pre-trained by human developers, it is ready to learn on its own as more queries come from users. An OpenAI study also found that about 19% of U.S. workers said they were affected in at least 50% of their jobs, suggesting that exposure to GPT is likely to be greater for higher-paying jobs, It is believed to be applicable to almost all industries. According to OpenAI’s research, these models will become general-purpose technologies “like steam engines and printing presses.”

Benefits of AI

What kind of work is AI good at?

A January 2023 paper by Anuj Kapoor of the Indian Institute of Management Ahmedabad and co-authors explored the question of whether AI tools or humans are more effective in helping people lose weight. The authors conducted the first causal assessment of the effectiveness of human and AI tools in helping consumers achieve health outcomes in real-world settings. This was done by comparing weight loss results achieved by mobile app users using AI alone. Some use coaches, others use human coaches.

Interestingly, while the human coach scored higher overall, users with higher BMI did not fare as well with the human coach as did those with lower body weight. ”

“Our findings may extend beyond the narrow domain of weight loss apps into the more general healthcare realm. Human coaches help consumers reach their weight loss goals. documented superiority to AI coaches in helping, and importantly, there are large differences in this effect across consumer groups, which suggests that a one-size-fits-all approach may not be the most effective. It suggests there may not be,” Kapoor told The Indian Express.

Findings: Human coaches help consumers reach their goals more than AI coaches for consumers below median BMI compared to consumers above median BMI. Human coaches help consumers reach their goals more than AI coaches for consumers below the median age compared to consumers above the median age.

Human coaches help consumers reach their goals better than AI coaches for consumers below the median duration compared to consumers spending more time than the median duration To do. Furthermore, human coaches help consumers reach their goals better than AI coaches for female consumers compared to male consumers.

Kapoor said the paper did not delve into the “why” of the effectiveness of AI+human planning for people with lower BMI than those with higher BMI, but he speculated on the reasons for the trend. ing. Feelings of guilt while dealing with other human beings. This is not always true, but generally there is good evidence to suggest this. Research has shown that individuals feel shame when purchasing birth control pills or consuming high-calorie indulgent foods. You may find it difficult to interact with other human coaches. This does not mean that health tech platforms should not propose human plans for people with high BMI. Instead, (1) train the coach enough to make individuals with high BMIs feel comfortable and be heard, and (2) guide him for weight loss. He determines the best combination of AI and human components. I can focus on what I do,” he added.

Similarly, the successful response of female consumers to human coaches can be attributed to recent advances in the literature on human-AI interaction. This suggests that AI adoption is different for women and men, and there are differences in adoption by age, she said Kapoor. This could be a potential reason for the difference in the impact of human coaches on men and women, they add.

In a previous OECD paper on AI and employment, “New Evidence from the Occupations Most Exposed to AI,” the impact of these tools “will be skewed in favor of highly skilled white-collar workers.” claims. manager; scientific and engineering expert; legal, social and cultural expert”.

This contrasts with the impact of previous automation techniques. Previous automation technologies tended to largely take over routine tasks performed by less skilled workers. A 2021 study said that increased exposure to AI “may be good for workers, as long as they have the skills to use these technologies effectively.” The study found that from 2012 to 2019, greater exposure to AI led to more employment in computer-intensive occupations. It suggests that workers with strong digital skills may be more capable of adapting and using AI in the workplace. So you can enjoy the benefits that these technologies bring. In contrast, there are some indications that greater exposure to AI leads to lower growth in average working hours in occupations with less computer use. Overall, the findings suggest that the adoption of AI “could widen the labor market divide between workers who have the skills to use AI effectively and those who do not.” I’m here. Ensuring workers have the right skills to work with new technologies is therefore a key policy issue that policy makers must increasingly address.

© Indian Express (P) Ltd