Technology could increase global GDP by 7%, but it also poses a risk of causing’significant disruption’According to Goldman Sachs research, artificial intelligence breakthroughs could allow for the automation of 25% of work in the US and eurozone.
According to the investment bank, “generative” AI systems like ChatGPT that can produce content indistinguishable with human output could trigger a productivity boom that will eventually lead to a 7 percent increase in global gross domestic product over a 10-year span.
However, if the technology lives up to its promises, it will also cause “significant disruption” in the labour market. According to Joseph Briggs, Devesh Kodyani, the paper’s writers, this would expose the equivalent of 300mn workers across large economies to automation. Lawyers and administrative personnel would be at highest risk of being made redundant.
Based on data from thousands of occupations, they calculate that AI automation is affecting approximately two-thirds (or more) of all jobs in America and Europe.
The majority of people would have less than half their work automated. They would likely continue to do their jobs and be able to use some of their time for more productive activities.
They calculated that this would apply to 63% of the US workforce. The remaining 30% who work in outdoor or physical jobs would not be affected, though their work may be vulnerable to other forms automation.
About 7 percent of US workers work in jobs that could be automated with generative AI. They are also vulnerable to being replaced.
According to Goldman, its research showed a similar effect in Europe. Globally, AI could replace manual jobs, which is a larger share of the workforce in developing countries. This means that AI could take over a fifth of all work, or 300 million full-time jobs in big economies.
This report will spark debate about the potential of AI technology to both revive the poor world’s stagnant productivity growth and create a new class dispossessed white-collar worker class. They could suffer a similar fate as the 1980s manufacturing workers.
Goldman’s estimates are less conservative than some academic studies that included effects from a wider range technology.
OpenAI, creator of GPT-4 and author of this paper, found that 80% of US workers could have at least 10% of their tasks done by generative AI based on the analysis of human researchers and the machine large language model (LLM) of the company.
Europol, the law enforcement agency warned this Week that rapid advances in AI could help online fraudsters and cybercriminals, so that “dark LLMs” may be possible. . . “This could be a major criminal business model in the future.”
Goldman stated that US investment could reach 1 percent of US GDP if AI investment continued to grow at the same pace as software investment in 1990s.
Based on analysis of data from the US and Europe on tasks performed in thousands of occupations, Goldman’s estimates were based. Researchers assumed that AI could perform tasks such as processing tax returns for small businesses, evaluating complex insurance claims, or documenting the findings of a crime scene investigation.
They didn’t envision AI being used for sensitive tasks like making a court decision, checking the status a patient in critical health or studying international tax laws.