Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Do you have an AI disclosure rule?

A recent poll shares how organizations are thinking about internal and external AI acknowledgements.

[Photo: Getty Images]

Profile photo of Sheri Jacobs

CEO, Author, Global Keynote Speaker · Avenue M Group

It’s Monday morning, and your inbox is piled up with emails that were clearly mostly written by AI. The poor, vilified em dash has nearly vanished, but these emails have all the hallmarks. A reply to a simple question runs four paragraphs and opens by thanking you for your continued partnership.

Each of the authors is simply trying to escape the quagmire of their own inbox a few minutes earlier. In doing so, they have shifted the burden to you, in the form of AI noise that swamps the message’s signal. We’re all getting sick of spending so much time in AI’s uncanny valley, where the appearance of substance and polish is not the fact of either.

Nobody disclosed anything about their use of AI or even considered whether they should. But the issue isn’t really if an AI tool was used, it’s whether the person on the receiving end has a right to know. After a few months of sorting through these emails on a fairly regular basis, I wanted to find out how leaders feel about this, so I asked them.

In an executive panel survey I conducted via text and LinkedIn last month, I asked, “Should employees have to disclose when they use AI on their work?” Of 124 respondents, 41% said no, it’s just a tool; 31% said yes, always; 18% said only for external work; and 10% said only if asked. A plurality sees no need for disclosure, but nearly half believe AI use should be revealed proactively at least some of the time.

A PUBLIC TEST OF THE QUESTION

The debate played out in public last August, when The Wall Street Journal published an op-ed by billionaire investor Stanley Druckenmiller challenging the Treasury Department’s intervention in the U.S. bond market. The conversation (riveting as it was) was hijacked by the revelation that most of the text was AI-generated.

Confronted, Druckenmiller was unapologetic. “Of course I used AI,” he said. He compared it to using a calculator and argued that what mattered was that the piece carried his name and his message. The Journal agreed, saying it considers whether a piece reflects the author’s own argument and whether the author has the credibility to make it. Other publications take the other side. The Financial Times has published post-hoc quasi-retractions of articles that it discovered had been authored with generous AI assistance. Every academic journal I’ve surveyed has staked a position mandating disclosure at the least.

One of the biggest challenges about disclosing AI assistance is the ever-increasing ubiquity of the technology. As AI worms its way into more and more of the software we use every day, the line between “written by me” and “written by me and CoPilot” gets harder to find.

PERCEPTIONS OF QUALITY

In the trade and professional association community, where I’ve worked in senior management roles and have conducted research for the past 17 years, members make the case that they pay for expertise and insights that carry the organization’s imprimatur. The notion that the content could have been generated without disclosure, especially by a publicly available LLM, weakens the brand for them. One survey respondent noted that their professional society requires a statement on any deliverable in which AI played a role.

Those who voted to mandate disclosure only for external work reflect how some leaders already weigh risk. One respondent commented that their organization uses the public-only disclosure model, with an honor system for internal work. Such a policy begs off on some important details, such as whether a board presentation counts as internal communications. Those who voted to mandate disclosure “only if asked” picked the most permissive option. Sniff if you want, but The Wall Street Journal is in their camp, at least for op-eds.

Of the 10 leaders who answered my follow-up about whether their organization has a rule, four described a formal policy, two are drafting one, and four said there is no rule or people decide for themselves. That mirrors a 2025 KPMG and University of Melbourne study of more than 48,000 people in 47 countries. The study found that only 40% of employees said their workplace has generative AI guidance, and 57% said they have hidden their AI use and presented AI-generated work as their own. One leader with a policy in place offered sound advice: A policy works only when it is easy to understand and communicated regularly.

HIDDEN AI USE AND WORK CULTURE

In a recent Harvard Business Review article, Jayshree Seth of 3M and Amy Edmondson of Harvard Business School describe “safety by proxy,” where employees circulate AI-generated concerns without names attached because speaking up in person feels risky. They describe AI as a workaround where people don’t feel safe. It is an equalizer for newer staff or those working in a second language, a rehearsal space for testing arguments, and a mirror that helps leaders hear objections their authority tends to filter out. The most valuable uses are also most likely to go underground if disclosure carries a stigma. A simple disclosure rule established by leadership can remove that guesswork.

The poll I conducted in September showed that leaders are split almost evenly on whether AI disclosure should be required. I think the bigger issue is that in most organizations, the decision is being made one employee at a time. When leaders don’t provide guidance, employees will use AI as they see fit, without disclosing and without much thought to the additional work they are creating for the reader.

Sheri Jacobs is a keynote speaker, author, and CEO of Avenue M Group.