MIT's Center for Information Systems Research just put a number on something a lot of directors have suspected but couldn't prove: boards that actually understand AI outperform.

Companies with AI-savvy boards posted return on equity 10.9 percentage points above their industry average. Companies without one lagged 3.8 points below. Roughly 15 points separate the two groups, and the only variable is how well the board itself understands the technology.

That gap didn't exist a few years ago, at least not in this form. By 2024, 72% of large-company boards had cleared the bar MIT set back in 2019 for "digitally savvy," up from just 24% five years earlier. But when MIT raised the bar to cover cutting-edge AI specifically, only 26% of boards cleared it. Most boards are fluent in a kind of tech literacy that no longer sets anyone apart and well behind on the one that does.

What "AI-savvy" actually means

MIT's researchers found that the boards pulling ahead organize their AI work around three things:

• Strategy - they can actually spot where AI is opening up new threats and opportunities in their industry;

• Defense - they're keeping pace with the cyber risk and compliance burden AI adds; and

• Oversight - they can tell whether the company's AI spend is creating real value and whether data is being used the way it should be.

None of that requires directors to become engineers. It requires two things most boards still don't have: personal, hands-on experience using AI themselves, and a clear, consolidated view of where AI is actually moving the needle in their business rather than three separate conversations about it happening in three separate committees.

Why the gap is likely to widen, not close

The uncomfortable part of MIT's finding is the trajectory. The 2019 "digitally savvy" bar took most boards five years to clear, and by the time they did, it had stopped moving the needle. AI is moving faster. A board that starts building fluency today is still working against an innovation curve that keeps steepening underneath it, which means the 15 point gap is more likely to widen than close on its own.

So what actually closes that gap?

Closing the gap in practice

Directors build fluency by using AI on real material, not by discussing it in the abstract. And a board only develops a real view of where AI is reshaping its industry if that picture sits somewhere directors actually look, alongside everything else they're already tracking, rather than in a separate briefing nobody opens twice.

That's the thinking behind how BoardRadar approaches AI for its own board users. Directors can work through board materials directly inside the platform with AI Analyst, building the kind of hands-on comfort MIT's research points to, without the unacceptable risk of pasting sensitive information into a personal AI account to get it. And because AI exposure and disruption risk show up alongside the rest of a board's intelligence, the same place directors already check competitive positioning, governance risk, and financial performance, strategy, defense, and oversight stop being three separate conversations and start being one continuous picture.

The boards MIT measured didn't get ahead by studying AI from a distance. They got ahead by using it and by seeing clearly where it mattered. That's a good description of what board stewardship should look like from here on out.