AI investments may not bring great returns in the near future: Goldman Sachs

All the time and money invested by various companies in artificial intelligence may provide disappointing returns, Goldman Sachs wrote in its recent report titled, ‘Gen AI: Too much to spend, too little benefit?’

Why is AI expensive?

"AI technology is exceptionally expensive, and to justify those costs, the technology must be able to solve complex problems, which it isn't designed to do," the report read, quoting Jim Covello, the head of Global Equity Research at Goldman Sachs.

"In our experience, even basic summarization tasks often yield illegible and nonsensical results," he said in the report.

Also Read: Gujarat government signs MoUs with IBM, Microsoft, and NASSCOM for AI development

Popular examples of this would be when Google’s Gemini AI began giving its users odd suggestions, including putting glue on pizza and depicting historically inaccurate imagery.

"The starting point for costs is also so high that even if costs decline, they would have to do so dramatically to make automating tasks with AI affordable," he added.

Why are AI investments so expensive?

Most of the money being invested into AI technology is actually spent on its infrastructure and hardware requirements. This includes data centers, the power grid, and chips.

How much can AI improve productivity?

AI will increase productivity in the U.S.by only 0.5% and promote economic growth by just 0.9% over the next decade, forecasted Daron Acemoglu, a professor of economics at the Massachusetts Institute of Technology.

Also Read: AI companies race to adapt chatbots to India’s many languages

Acemoglu feels that there will not be a "massive" number of jobs impacted by AI in the near future as he says that a lot of jobs such as those in manufacturing or mining are "multifaceted and require real-world interaction."

He added that it would impact "pure mental tasks" more, but the amount of these jobs will not actually be huge.

Also Read: India needs an extra 1.7-3.6 GW data centre capacity by 2028 due to high demand

However, not all experts quoted in the report are pessimistic, with some telling that while AI technology is expensive today, it will become more affordable in the future, just like any piece of technology in general.