Please Stop Inviting AI Notetakers to Meetings

Using artificial intelligence to summarize meetings raises questions around etiquette, privacy and the purpose of meeting in the first place.

See omnystudio.com/listener for privacy information.

Episode Transcript

Speaker 1 (00:00):
AI note takers are making me queasy by Chris stokal
Walker read by Mark Lee move over. This meeting could
have been an email The new oafas complained, could soon
be who invited the AI note taker. Microsoft Corp, Google,
and a bunch of much smaller firms have all rolled
out artificial intelligence powered note taking tools in recent months.

(00:24):
These bots listen in on meetings, transcribe what was said,
and summarize key points. The tech companies behind this shift
might frame their use as a step forward and office efficiency,
but the technology raises troubling questions around etiquette and privacy,
and risks undercutting the very communication it is meant to improve.
Google Rajaram, co founder of Marathon Management Partners, a San

(00:46):
Francisco tech investment firm, says there are AI note takers
at eighty percent of the meetings he attends these days.
Sometimes two or more bots brought in by different attendees
will be transcribing and summarizing at the same time. The
bots are constantly improving their abilities to synthesize what was discussed.
Roger Ram says, for people who have not attended the meeting,

(01:08):
it's a great way to understand what went on even
for those who are in the meeting, it now saves
you from taking notes. But he's also concerned about the
potential for AI to hallucinate quotes based on its training data,
or to extrapolate what was said into something that wasn't,
and about what happens to information after the meeting. It's
not just taking notes, says Rojeram. It's recording your voice

(01:32):
that hits on the biggest issue with its new business
ethics conundrum. When I was first confronted with a clutch
of AI note takers in a Zoom call last year,
I was taken aback. My inputs of the conversation changed,
knowing it would be recorded and summarized by AI. It
felt invasive and like I had lost control over what
would happen to that information. My words could be taken

(01:54):
out of context and used against me. My thoughts and
my words could be used as training data by tech
company I had never heard of. Mostly these note takers
are turning up at virtual meetings, but they can be
prison in real life too, via laptop or a phone.
Using an AI note taker is much more like recording
a meeting than taking your own notes, says Margaret Mitchell,

(02:15):
chief ethic scientists at Hugging Face and AI Company. Some
of the etiquette around recording applies, she says. For instance,
anyone using an AI not taker should let others in
the meeting know and give them the right to veto
the AI's attendance. Mitchell says the rise of the note
takers also touches on some big tech issues. As we
seek to have more automation and traditionally human tasks, we

(02:38):
seed our control, power and privacy to AI systems and
the companies that own them. She says she's worried about
the limitations of the bots too. A critical thing that
AI not taking lacks, which human note takers have are
the nonverbal cues from the speakers, she says, are sarcastic
joke said in a meeting could become an action point

(02:59):
when crank through the acute power of an AI system.
The tech booster answer to that problem might be to
get AI note takers to incorporate visual processing as well
as verbal. But then we're at the surveillance issue, she says,
and your harried Monday morning visage becomes another data point
to count against you. The push to document every workplace

(03:19):
interaction and utterance is not new. Having a paper trail
has long been seen as a useful thing, and a
record of decisions and action points is arguably what makes
a meeting meaningful. The difference now is the inclusion of
new technology that lacks the nuance and depth of understanding
inherent to human interaction in a meeting room. In some ways,

(03:40):
the prior generation of communication tools such as instant messaging
service Slack created its own set of problems. Messaging that
previously passed in private via email became much more transparent,
creating a minefield where one wrong word or badly chosen
emoji can explode into a dispute between colleagues. There is
a similars with note taking tools. Each utterance documented and

(04:03):
analyzed by AI includes the potential for misteps and misunderstandings.
Anyone thinking of bringing an AI no taker to a
meeting must consider how other attendees will respond, says Andrew Brodsky,
Assistant Professor of Management at Macomb's School of Business, part
of the University of Texas at Austin. Colleagues might think
you want a better focus on what is said without

(04:25):
missing out on a definitive record of the discussion, or
they might think you can't be bothered to take notes
yourself or remember what was being talked about. He says,
for the companies that sell these AI interlopers, the upside
is clear. They recognize we're easily nudged into different behaviors
and can quickly become reliant on tools that we survived
without for years. The AI transcription service Otter happily drains

(04:49):
two hundred and forty dollars from my bank account every year.
Despite the limited efficacy of Otter's AI summaries. The two
action items from its analysis of my call with Bronsky
reach out to the edit to send over the article,
and feel free to reach out to Andrew any other
time for future quotes or discussions. And yet, if you
ask me whether I'd want to transcribe all my interviews

(05:10):
by hand, I'll laugh at you, even though that's exactly
what I did for years before Otter arrived. There's another
benefit for tech companies getting us hooked on AI note takers.
Training data for AI systems is increasingly hard to come by.
Epoch AI our research group forecast there will be a
drought of usable texts, possibly by next year and with

(05:31):
publishers unleashing lawsuits against AI companies for hoovering up their
content to train the tech firms are on the hunt
for other sources of data. Notes from millions of meetings
around the world could be an ideal option for those
of us who are the source of such data. However,
the situation is more nuanced. The key question is whether
AI note takers make office meetings more useless than so

(05:52):
many already are. There's an argument that meetings are an
important excuse for workers to come together and talk as
human beings. All that small talk is where good ideas
often germinate. That's ostensibly why so many companies are demanding
staff return to the office. But if workers trade in
person engagement for an AI summarized read back, and colleagues

(06:13):
curb their words and their ideas for fear of being
exposed by the bots, what's left in the end, If
the humans step back, all that remains is a series
of data points and more AI slop polluting our lives.