Leaders are leaning on AI to guide their actions, with few clear benefits so far. It may help to see AI as part of a process, not its end.

Two recent data points about AI, both troubling:

One: According to a recent PwC CEO survey, only 12 percent of organizations have increased revenue and decreased costs through the use of AI.

Two: A survey of UK executives by the tech firm Confluent found that “62 percent of executives use AI to make the majority of their decisions, while only 18 percent report not using it at all.”

These two things aren’t strictly correlated, I recognize. Leaders offloading their decisionmaking to bots doesn’t necessarily affect the bottom line, and there may be benefits to using AI that aren’t strictly financial. The Confluent report makes that case, suggesting that leaders still rely too often on “trust your gut” decision-making, and that the faster pace of doing business means leaders need to act faster in a data-driven way.

But here’s the sticky part: Leaders’ use of AI seems to encourage them to behave as if they don’t have colleagues, or need them. According to the Confluent report, “65 percent feel that decision-making has become less collaborative as a result. Almost half (46 percent) of bosses say they now rely more on AI than their colleagues’ advice to make decisions.”

If associations are the collaborative organizations they assert themselves to be—steered intelligently by diverse and informed boards, as they often claim to be—over-reliance on AI for direction-setting can only be bad news.

At the very least, it suggests that a lot of the talk that CEOs deliver about innovation and fresh thinking and collaboration is emptier than one might hope. It also suggests that now that many people are outside of panic mode when it comes to AI, things have now swung in the other direction, where organizations are using AI in a try-anything manner, still searching for clarity about where it can and can’t assist, and helping their people better understand it.

The PwC report makes that point: “While many workers have experimented with AI, only a small minority – 14 percent – are using it daily at work. Enthusiasm is high, but clarity and enablement lag. Enterprise value depends on translating AI ambition into redesigned work, skills and day-to-day practice.”

A helpful section in Deloitte’s recent hefty Global Human Capital Trends report both highlights the scope of the problem while proposing some solutions. Its CEO survey found that while 64 percent of leaders recognize the importance of AI’s implications around decision making, only 5 percent are making “great progress” around that.

So now would be a good time to start progressing. The Deloitte report highlights the importance of designing decision frameworks, where organizations collectively determine which decisions are low-risk enough to move forward with confidently, and which require more collaboration and scrutiny.

Just as important, organizations need to determine where and how AI fits in those processes. “With AI, [decision] rights need to be more dynamic, incorporating override privileges, escalation paths, and consensus rules engineered into the system so humans and agents coordinate who decides, when, and on what basis,” the report says. 

None of this is intended to slow interest in AI as a tool—no question, it can gather and summarize relevant data in powerful ways, and associations would be as mistaken to avoid using it, just as they would be mistaken for passing on social media in 2008. But ultimately, human beings have to be in charge. That means deciding how AI will assist human decision-makers, not replace them.

The post Can AI Make Executive Decisions? appeared first on Associations Now.