As corporate spending on artificial intelligence systems is set to pass $50 billion this year, the vast majority of companies may not be seeing much return on that record investment.
In a survey of more than 3,000 company managers about their AI spend, only 10% reported significant financial benefits from their investment so far, the new report from MIT Sloan Management Review and Boston Consulting Group found.
Gains from the tech haven’t kept pace with increased adoption, says Shervin Khodabandeh, who led the study and is co-head of BCG’s AI business in North America. “We are seeing more activity, which also means more investment in technology and data science,” Khodabandeh says. “But that impact line hasn’t really changed.”
The results should prove concerning to corporations that continue to pour money into AI projects at a breakneck clip, looking to use the tools for everything from managing contracts to powering home assistants and self-driving cars. More than $50 billion is expected to be invested in AI systems globally this year, according to IDC, up from $37.5 billion in 2019. By 2024, investment is expected to reach $110 billion, IDC forecasts.
But despite the billions invested, failed AI projects have become an increasing factor. IBM has deprioritized its Watson technology after drawing scorn for ventures like one $62 million oncology project that made inaccurate suggestions on cancer treatments. Amazon canned an AI recruitment tool after it showed misogynistic biases. And smaller businesses have found that building the technology is harder than it looks, as supposedly AI-powered virtual assistants and meetings schedulers end up relying on actual humans behind the scenes.
Companies are struggling to deliver on AI projects, Khodabandeh says, because they overspend on technology and data scientists, without implementing changes in the business processes that could benefit from AI — a conclusion that echoes a Harvard Business Review report published in June.
Take Uber. Last month, engineers at the ride-hailing company concluded that its self-driving cars couldn’t drive more than half a mile before encountering a problem. The program’s artificial intelligence still “struggles with simple routines and simple maneuvers,” per a report in The Information. Part of the reason for the failure, according to an internal memo: competing internal ideas on how to implement the tech.
But with AI’s promise of large-scale business savings and improvements, companies aren’t likely to stop investing in the technology soon. The BCG and MIT researchers found that 57% of companies said they’ve deployed or piloted their own AI projects, up from 44% in 2018.
For those projects to pay off, Khodabandeh says more AI adopters will need to rethink how the tech is integrated within their businesses. “There’s clearly a lot of hype,” he says. “And some of that hype comes out in the data.”