When every board member gets the same AI summary, who has done the thinking?

AI summaries may make a board pack quicker to read while making directors’ advice more alike. In a Forbes essay, AI executive Ajay Pundhir argues that summaries should point back to their sources, and that directors should be assigned the sections behind important decisions.

Watch Desk analysis

What happened

Pundhir draws on a study by Shiri Melumad and Jin Ho Yun, which he says involved seven experiments with 10,462 people. In one experiment, people who received an AI synthesis wrote shorter advice with fewer distinct facts than people who searched several sources containing the same information. Their advice also became more similar to one another’s.

He contrasts that with a Harvard Business School working paper on summaries attached to Wall Street Journal articles. As Pundhir describes it, readers were more likely to search afterwards, and students who saw a summary alongside the full article remembered more. The distinction is whether a summary sits beside the source or stands in for it.

His proposed “index rule” is simple: assign named sections of the underlying document to named directors, and require each point in the AI summary to link back to the page or paragraph it compresses. Pundhir presents this as a practical proposal, not a boardroom-tested result.

Why it matters

If everyone starts with the same compressed account, a board can mistake one interpretation repeated nine times for nine independent judgements. Pundhir’s answer is not to ban summaries, but to make them useful signposts and ensure someone has read the evidence behind consequential points.

That matters well beyond boardrooms. AI-generated digests can save time, but readers still need a route back to the material that supports a claim, especially when a decision turns on details the digest may have left out.

Our read

The strongest point here is not that summaries are inherently bad. It is that convenience can quietly become substitution. A summary with references invites a reader to check; one without them asks the reader to take its framing on trust. That is a rather large job for a small box of text.

Pundhir’s index rule is a sensible proposal, though the cited experiments were not tests of corporate boards. Treat it as a practical prompt for board chairs, not a proven governance fix.

What to watch

• Whether board packs link AI-generated claims to the original pages or passages.
• Whether directors are assigned the evidence behind decisions that matter.
• Whether further research tests these practices in real boardrooms.

Discussion spark: For a consequential board decision, should every director read the same full material, or is it better to divide the reading and make the AI summary point back to the evidence?

Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.