Named 2018 PR Firm of the Year by PRNews

3 Things to Consider Before Using an AI Note-taking Service

posted by:

Jenna Sarkissian

18 July

If you’re in a business of any kind, you’re no stranger to meetings — and if your business is public relations, lots of meetings. You may be on the lookout for tools to increase productivity and focus, and something like an artificial intelligence-based notetaker might catch your eye — it certainly piqued my interest!

It can work something like this: You sign up for the service online, send it a calendar invite to your meeting, and then forget about the rest. The AI-based notetaker calls in, takes notes, then sends you a meeting summary right after, with audio clips included.

Though several new and exciting artificial intelligence-based note-taking companies like Voicera and Clarke.ai boast how their services can help you increase productivity, I challenge eager professionals to take a step back first and consider some ethical and legal dilemmas that may arise from using such services.

Here are three questions to consider before you start using an AI-based note-taking service:

1) Do I live in a two-party consent state?

U.S. map showing where AI note-taking needs two-party consent

Photo courtesy of: Diligentia Group

In 12 states — California, Connecticut, Florida, Illinois, Maryland, Massachusetts, Michigan, Montana, Nevada, New Hampshire, Pennsylvania, and Washington — all parties involved in a meeting where there is a reasonable expectation of privacy must give consent before the conversation may be recorded.

With some AI note-taking tools, all you have to do is add an email address to your calendar invite, then the AI-notetaker will automatically record and transcribe the meeting you invite it to. But merely adding the AI notetaker to your calendar invite does not serve as appropriate notice to the parties being recorded if you live in one of those two-party consent states, said Clay Calvert, an attorney and director of the Marion B. Brechner First Amendment Project at the University of Florida.

Headshot of Clay Calvert, expert on AI note-taking issues

Photo courtesy of: UF College of Journalism and Communications

“This is a wake-up call for most people because they wouldn’t think about these types of issues,” Calvert said. “But people on the other end of the conversation have a reasonable expectation of privacy in this case, especially if there has not been a history of recording meetings.”

2) Have I notified the meeting participants and obtained proper consent?

One company, Clarke.ai, offers these best practices for using its service in all states, which includes disclosing that you are using an artificial intelligence notetaker. But in two-party consent states, you should take it a step further.

Before the call, you’ll want to get each person’s consent before you start recording them, Calvert says.Then once the recording begins, you should again ask for approval so you’ve documented it as part of the transcript.

A statement like “Just to reconfirm, did you give consent for my AI-based note-taking tool to record this call?” should work just fine.

You should also consider with great care your company’s policies on recording meetings, and then be sure to follow those policies. In my opinion, it’s always better to make clear to everyone involved that you’d like to use the service well in advance, so there’s ample time to discuss it and ensure that everyone is comfortable.

3) Have I done my research and considered the risks?

Most of the AI-based note-taking services that I found through a quick Google search online are relatively young — both companies mentioned above were founded in 2016, according to crunchbase.com. That’s not a lot of time to make mistakes and learn from them — or for laws and regulations to catch up.

It may seem silly to take such a cautious approach to seemingly harmless AI-based note-taking tools, especially as AI touches many areas of our lives — voice-based devices like Amazon’s Alexa and Google Home devices are seemingly everywhere. Still, I stand by my recommendation to tread cautiously, especially as you incorporate AI into your work life and business practices where there’s likely to be sensitive and confidential information shared.

Some questions that remain: Where is meeting data stored? Who has access to the data? How long is the data stored?

cell phone screen showing dashboard for AI note-taking service, Voicera

Photo courtesy of: VOICERA and Forbes.com

I recently posted one of these questions on a Facebook thread. Voicera responded:

“Nobody at Voicera listens to your meetings. The transcription is all AI driven and they are stored in the cloud where you enable access only to those who you designate. Our goal is for your meetings to be yours and only yours.”

Still, I believe the privacy implications are quite real, not only for this kind of tool but in all forms of AI. Harvard University’s Berkman Klein Center for Internet & Society shares a lot of compelling content on this topic, such as this recent video on Twitter. Whether you’re an engineer or a writer, everyone has a role to play in understanding the impact and ethics of AI before jumping on board.