One of our favorite and most colorful clients starts every video call with the prelude: “Everybody on this call, turn off the damn AI notes software!”
Despite the world’s exuberance with artificial intelligence, here’s why I agree: As video calls pervade the workplace, it seems every video call gadget is trying to start up an AI tool in the background to transcribe every single word and later send you a helpful “Summary of Key Points.”
Since I’m a former tech reporter, you might think I would love AI innovation here. Nope. I’ve polled specialists in privacy, management and law, and they have a universal reaction to AI note-taking software: “Turn it off! Now!”
Rich Mullins
Here’s why:
- First, you should realize AI may already be listening; it is often embedded in video platforms. It can appear as a meeting attendee with cute names like “Firefly Fred,” or “Industrious Squirrel.” Where do these AI notes go? How easily can that transcript leak? Do you like the idea of emailed copies of transcripts floating around?
- Privacy matters and always has. I suspect ancient Egyptians had feisty staff meetings while building pyramids. Now businesses hold video meetings about research projects, competitors, wild rumors and problem employees. These should all be private.
- The law matters. Each state has a variety of laws about recording or monitoring conversations. Some require overt consent to record conversations. So, what about a call with seven people in six states? Also, any call that includes a government employee can easily make that transcript into a public record.
- Casual banter kills. In our media and public speaking training for CEOs, we suggest they avoid jokes. Humor rarely translates well. But alas, something about video calls spurs people to joke around, and the AI software could transcribe everything as fact. “Has the CEO been drinking again? Har, har, har?” Try explaining that transcript to the CEO, or a jury.
- Cybersecurity. I may have more implicit trust in Google or Apple, but do you trust a six-month-old startup to listen to your company’s most confidential material? What if hackers stole transcripts of calls at hospitals, military sites or reporter interviews with confidential sources? What if hackers got transcripts of Justin Baldoni’s calls with his lawyer about suing Blake Lively? Spicy stuff, I bet.
I will add that one AI transcription company boasts on its website that it “generates your meeting notes in 30 languages, auto-updates your CRM, drafts your next follow-up email and sends AI-generated summaries of the last 20 meetings.” Dare I say that sounds invasive. And their service is “Free, forever.”
To paraphrase Apple CEO Tim Cook: “When a service is free, you’re not the customer. You’re the product.”
What to do, then? HR specialists and lawyers I polled suggest this: First, inventory what AI transcription systems your staff already uses. Second, ensure those programs match your company security policies. Third, establish rules and training for using or blocking transcription tools. (Have strict rules about staff using personal devices on calls too.) Fourth, make sure someone in your organization knows federal and state privacy laws.
To be sure, I suspect plenty of AI transcription companies have good counterarguments, and there could be plenty of places where AI notetaking is a Godsend. College, for instance. I would have paid dearly for software to transcribe my Middle East history classes. Also, a savvy litigator friend of mine ‘likes’ AI transcription for calls with opposing lawyers so he can insert advantageous evidence into the record. And AI summaries may be useful to companies to summarize already-public calls, like earnings announcements. So, perhaps it’s best to remind ourselves that any new technology is just a tool. It’s powerful when used the right way. Dangerous when not.
For now, I agree with our most colorful client: when in doubt, turn the AI off.
Richard Mullins is a vice president at Tampa-based Tucker/Hall Inc., specializing in crisis and strategic communications.