The colossal increase in investment in artificial intelligence may still take a few years to appear significantly in American GDP statistics, but recent academic studies already show gains of 25% on average (and 16% on the median) in worker productivity.
These are some of the conclusions from Goldman Sachs, where the macroeconomics team took stock of the pace of technology adoption in the US and its effects on the country’s economy so far.
A year ago, Goldman estimated the possible increase in labor productivity in the US over a ten-year period at 15%. After 12 months, analysts say that the transformation in the economy is underway, but that the benefits are still concentrated in a restricted number of players and sectors, the early adopters.
Less than 5% of US companies are formally using generative artificial intelligence tools. The percentage of adoption, however, is 2x or 3x higher in sectors such as finance and information technology.
In business surveys, executives list a series of reasons that hinder the faster adoption of AI, including a lack of knowledge of how to apply the technology and concerns about privacy. For Goldman, these barriers should lose strength in the coming years.
So far, there is no evidence of major effects with regard to the feared destruction of jobs in the economy. On the contrary.
“Low adoption has limited the impact on the labor market, but preliminary evidence suggests that AI is modestly increasing demand for people and has led to negligible job losses, therefore creating a slight positive boost in net hiring,” say analysts at Goldman.
You early adoptershowever, achieved a “great increase in worker productivity.”
Initial estimates should be interpreted with caution, says Goldman, but academic studies have indicated an average increase of 25% and a median of 16% in company productivity after incorporating AI into their activities.
“Anecdotal reports from companies similarly suggest large efficiency gains,” says the bank.
Most investments have focused on hardware – chips and data centers. According to Goldman, it should reach US$250 billion next year – an amount equivalent to 9% of all investments expected to be made in the USA.
“The considerable increase in investments in AI and the large productivity gains among early adopters support our confidence that generative AI represents a upside significant for the economy, but the slow pace of adoption suggests that the most relevant macroeconomic impact will still take a few years,” says Goldman.
Regarding the impact on the value of companies, Goldman notes that the valuations were concentrated in hardware companies (for example, semiconductor manufacturers) and cloud companies. It was a reflection of the increase in demand for AI chips and data centers. “The likely beneficiaries of productivity gains have seen modest gains,” the report says.
Semiconductor companies have seen their revenue increase by 50% since the beginning of 2023 – mainly because of the 200% jump in Nvidia’s quarterly sales.
Giuliano Guandalini
