more on AI attending meetings

by Alison Green on December 6, 2024

A reader writes:

This may be cheating as an “update” since I am not the original person who wrote in about AI attending meetings, but I still thought it might be interesting/useful to you and your readers!

Today I was in a Zoom meeting: a regular small meeting to publish a paper. I noticed an AI notetaker logged in as a bot for one of the meeting regulars; this was the first time I’ve seen one in the wild. But because I read your blog and had seen a question about AI note takers as well as the follow-up, I knew that the person may have been totally unaware about it. The bot actually logged on BEFORE the person logged on themselves. The person apologized that they were having some computer issues. I asked if the bot was intentional as I had heard these have started joining Zooms with the “owner” unaware. Indeed she was not aware and asked me to get rid of it.

Once I booted it, I got two emails from the org that generated the bot. One was a summary of what we talked about before I booted the bot. It has a link to a “full summary” but you have to make an account and log in to see it. The person said DON’T DO IT. It turns out that’s how she got “infected” by the bot. She didn’t want it and it followed her from her work computer to her personal one. Presumably it got some sort of access to her zoom account, and she didn’t intend any of that!

So just an FYI: while these bots actually do seem to provide an interesting summary of the meeting, they also seem to be propagating almost like phishing and probably everyone should avoid logging in to any websites directed by “summary” emails they receive unless they too want a pseudo-viral bot.

You may also like:
{ 97 comments }

{ 97 comments… read them below }

  1. ArchivesPony*

    We just had this come up in our library (academia) leadership team meeting. My institution is working on ethics for it and especially because we do have comply with our state’s open records act, things are getting dicey and weird

    1. Academic Hellscape*

      Transcripts in FOIA requests for meetings people didn’t know they were being recorded on is a nightmare waiting to happen. I also work in academia at a state institution, and support transparency etc. But I cannot imagine trying to do my work with no one willing to really talk about an issue for fear of a phase being taken out of context and on the front page of the local paper.

      1. Emily (not a robot)*

        Yes — people need to be able to have conversations with coworkers that don’t wind up becoming public records! And they still will even if meetings start being recorded — you just push them into saying things in private, in-person conversations, which reduces transparency because now fewer people in the organization even know what’s going on and why.

      2. Strive to Excel*

        I was in a prior job where we saw a lot of sensitive employee data, like, all of it, and our accepted way to handle situations where we needed to see it and confirm that we saw it for remote clients but didn’t want to permanently record it was via Zoom meeting. We would handle it via a note that says “auditor reviewed records and agreed the SSN from the employee record to the 3rd party 401k benefits provider w/o/e, SSN redacted for privacy reasons”. An AI transcription software doing that would be NIGHTMARISH.

    2. Banana Pyjamas*

      UNESCO published guidelines for the ethical implementation of AI. If you Google “unesco ethics ai”, it will be the first result. White House. gov has a “Blueprint for an AI bill of rights”, which is the first result when you Google “white house ai bill of rights”. The US Department of Labor also has guidelines on “Artificial Intelligence and Worker Well-being” which will come up if you Google “usdol ai best practices”. UNESCO and USDOL will probably be most helpful.

    3. Ann*

      That adds an extra wrinkle to the issue! One of my coworkers had a similar problem with an AI bot that went where it wasn’t asked. On top of that, on reviewing the transcripts, it turned out there were quite a few errors, like mis-recording things or not noting correctly who is speaking. And AI is mysteriously prone to making things up if it thinks there’s a gap in its knowledge. So making AI recordings part of public record could make things public that were never actually said at all…

      1. Person*

        We’ve had this come up in academia with AI bots attending sessions students have booked into for academic advice – sometimes without the actual student attending, which is super weird!

        I had previously thought the students were sending the bots in deliberately as some kind of misguided attempt to do two things at once, but with what the LW said I’m now wondering whether it’s actually just a normal no-show and the students don’t realise the bot is attending.

    4. Gabi Howard*

      I do IT for a number of charities, and we’ve now switched off *all* third party add-ins for *all* employees at several of them just because of how pervasive and hard to remove these things are.

  2. online millenial*

    Can’t say I’m surprised–most of the genAI tools run on opt-out, consent-free models. Force their way into your software, browsers, and machines so their numbers look better and they can claim to be popular and inevitable. Thank you for sharing this info!

    2. Antilles*

      That, plus they also collect a bunch of data they can harvest, add to their own databases, use for their own training, store it in their own completely safe* locations and sell it.

      *Safety not guaranteed, we assume no liability for data breaches.

      2. Bilateralrope*

        It sounds like it’s just a matter of time before one of them gets squashed due to GDPR violations.

    3. Zeus*

      I think it’s high time we start calling these “helpful” “bots” out for what they really are – spyware.

  3. UnionGal*

    My last place also had an issue like this, because the AI notetaker was listening in on Union grievance meetings and then HR was sending summaries to people who shouldn’t be having them! Once the union found out, AI note takers were SWIFTLY banned from the institution.

    2. Wired Wolf*

      Hmm… This is good to know; my workplace has a contract negotiation coming up. I wouldn’t expect any of my managers to understand why the AI notetakers are bad news (or even know how to detect one if it did worm its way in).

      One of the new learning modules claims to be about “ethical AI use within the company”, but the module itself is pathetic (little research done and the only takeaway is “know when to use it”).

      1. Gabi Howard*

        I’m willing to bet it was at least part-written by an AI company, or at least based heavily on their ‘educational’ promo materials.

  4. Keyboard Cowboy*

    That’s super wild – and super shady. Who else has access to the content and summaries of all these confidential work meetings? Who is paying for the compute for the AI model? This feels somewhere between phishing and corporate espionage to me, honestly.

    1. Bird names*

      At this point I wouldn’t be surprised if genuine legal action will need to be taken soonish if these tools keep proliferating like that.
      I’m not even in an industry with super-high privacy needs and I still suspect that about 90% of digital work conversations would be an issue with this kind of unasked-for surveillance and the remaining 10% would probably be covered by personal annoyance.

    2. Not a Girl Boss*

      “If the service is free, you are the product”

      I work somewhere that takes data security really seriously, and this sounds like a nightmare.

      1. LizardOfOz*

        Sadly, quite often even if the service is paid, you’re still the product.

        A lot of companies have now realized there’s a (quite profitable) revenue stream they weren’t tapping into (their customer’s data) and will happily sell that to whoever’s willing to pay.

    3. I Spend All My Nights At The Museum*

      Knowingly letting an AI collecting who knows what FOUO information into a government IT system is a recipe for loss of security clearance (I am a vet and used to work in military accounting). It sounds like something that could be used for ACTUAL espionage

  5. learnedthehardway*

    Wow! I’d say this is really sus (as my teen would put it), but that would be attributing intention/motive to an AI. (Mind you, maybe they do have intention, at this point).

      1. Pastor Petty Labelle*

        Someone created the bot and set it up to proliferate. Which is super shady and needs to be shut down immediately.

        1. DJ Abbott*

          This is akin to the idea that corporations are entities in themselves, and no individual person is accountable for what they do.
          Corporations are run by people, people make the decisions, they should be held accountable.
          AI was created by people, people programmed it to do this, they should be held accountable.

    2. RagingADHD*

      The decision to build them in such a way that they automatically attach themselves to your Zoom account without an explicit opt-in, and to make the opt-out very subtle or not display it at all, was made by humans. And very much with intent to get embedded in as many places as possible, without regard for privacy concerns.

      If we hadn’t already had the Cambridge Analytica scandal with Facebook secretly mining user data so they could sell it to the highest bidder, I would probably casually assume that nobody would be interested in sorting through a bunch of random transcripts or have a financial motive for doing so. But here we are.

    3. Brandon*

      An intention/motive may not have been a factor.

      This is a “convenience” feature that can be used for malfeasance.

      In my time in the space of rolling out software, every “bad outcome” is not vetted to its conclusion. Whether or not that is right or wrong is a different question.

  6. mango chiffon*

    I hate these AI notetaker bots. We sometimes have external advisors who sign non-disclosure agreements with our org because we will often discuss sensitive research data in meetings that we set up. One time, one of the advisors had an AI notetaker bot, and it really worries me. The bot is not privy to our NDA, this data is going who knows where to someone we don’t know. There is no consent to have these attend the meetings and it just showed up and took notes about our work. I had to bring this up to our legal department because they just updated our data privacy policy about our own use of AI bots, but there wasn’t anything about if external people have them. It’s such a wild west right now and I don’t trust anything.

    2. Kyrielle*

      I would think bringing in a third-party AI notetaker bot would be a breach of the NDA, but I also think that if it is, there’s still the problem that the person “with” the bot may not have meant to have it.

    3. Anon in Aotearoa*

      Add a requirement in your NDA that AI in the form of note taking bots or generative AI is not allowed. I am a freelance consultant and just signed a client NDA with wording to that effect. (I was happy to sign it as I don’t use AI).

  7. Ann O'Nemity*

    My org recently started using Zoom’s built-in AI Companion and is phasing out the use of the third party tools. We’re not banning them yet, but it’s heading in that direction. Who knows what the third parties are doing with the data they’re collecting from all these meetings. (I do think they’re making some exceptions for the legit accessibility transcribers that have been around for a long time in cases where the accessibility tool offers features that Zoom’s AI does not.)

    1. ArchivesPony*

      Just an FYI, Zoom’s AI, as of now, according to my library systems person will “learn” from stuff it’s turned on. If you don’t want that, they (our systems person) suggested either Gemini (google) or Copilot (Microsoft). It might be different for non-academic places but I’d look into that.

      1. Ann O'Nemity*

        According to recent updates and terms of service from Zoom, the company explicitly states that they do not use any customer data, including audio, video, chat, or screen sharing content, to train their AI models. But who knows; they’ve been slapped by the FTC before over misleading and misrepresenting business practices.

        1. Madame Desmortes*

          I would check also for clauses that say “we can sell your data to / share your data with our Special Friends (who may or may not use it to train their AI models)” as well. Common dodge.

    2. Zarniwoop*

      It would make sense to have a “whitelist” of trusted vetted programs, and ban anything else as a potential risk.

  8. Anonymous worm*

    What are you seeing? I’m only infrequently (once a month) Zoom meetings. How do you spot the AI note taker?

    1. oaktree*

      They might appear as an extra participant. Let’s say you invite Ernesto to a meeting. Your Zoom attendees might also list something like “Ernesto’s AI notetaker.” With or without Ernesto’s actual attendance.

    2. The Ginger Ginger*

      it actually looks similar to an attendee and will have it’s own attendee tile with its name (ex: GenericBrandAI Notetaker) and the employee owner (usually in parenthesis). You can’t miss them if you’re looking at your attendees.

    3. RagingADHD*

      Sometimes the bot attends under a brand name like Jamie, Leexi, Bubbles, Rev, Otter, Fathom, Fellow, Granola, etc. If you aren’t looking for an email address or other information, you might assume it was a person’s nickname, or an account that just had an unusual title. You should always check your attendee list for email addresses you didn’t invite and / or that don’t line up with the intended organization.

    4. Ann O'Nemity*

      Most of the third party tools show up like a meeting participant. They don’t ask for permission to take notes or transcribe, they just lurk there. Thankfully they tend to have obvious “AI” names that are easily identifiable. You can refuse to admit them if you have waiting room turned on, or boot them. I’m not a fan of the bots being a “participant,” though, since it can get weird with breakouts – you can get stuck in a breakout with multiple bots instead of real people if the organizers aren’t careful.

      Zoom’s AI Companion, on the other hand, is built in so it’s not another participant. All real participants get a warning when the AI is turned on, just like the recording feature.

    5. AI surprise*

      This was my update (woo!) – I was just so amazed it happened and the prior AAM post left me at least a little prepared to ask about it and boot it; it logged in like a zoom attendee, so you could see it there, and it was logged in as “[name]’s AI Notetaker (Otter.ai) has joined your meeting” – it looks like Otter.ai is the 3rd party app/bot/[object reference uncertain].
      When that person logged on they had their own account as usual and I was able to boot the bot without logging out the person who was supposed to be there.

      I then got an email (so it had access somehow to the email accounts from the zoom – which I may add I used my university account to set up the meeting, it wasn’t the other person’s meeting room)

      Email I received (with parts redacted)
      —————————–
      [title of meeting]
      [date of meeting]

      In the conversation, [yours truly] discusses the presence of an AI note-taker in a Zoom meeting, expressing curiosity and skepticism about its purpose and whether the participants are aware of it. [Other attendee] mentions editing a reference part and uploading a cover letter draft. The conversation also touches on personal topics, including [Other attendee]’s location and a brief mention of martial arts. Additionally, there is a discussion about a computer issue related to note-taking software, with [AI infected attendee] requesting its removal. The tone of the conversation is casual, with elements of humor and small talk.

      See full summary → (link to the sus AI tool sign in)

      Action items:
      Remove the AI note-taker from the meeting. (this is a link of some kind; I did not click on it; nt sure if it’s reporting the action item we talked about before it was booted or if it’s some kind of separate offer that just happened to be after the “action items” header)

      [name] is automating meeting notes. Try it for your own meetings!

      Get started for free → (link to the sus AI tool sign in)

      ———–
      FWIW that is fairly accurate except we briefly talked about the “martial law” episode in south korea, where one of the participants are, not “martial arts”

    6. Lenora Rose*

      They sometimes show up *only* in the chat menu, too, usually as “So-and-so”‘s bot, so presumably piggybacking on their feed.

  9. Yikes*

    Also not the original question-asker, but Allison’s advice was super helpful for writing my organization’s AI notetaker policy!

  10. Blarg*

    Verbatim summary, which I screenshotted because it was so funny.

    “Speaker 1 discusses the unexpected and sometimes intrusive nature of AI note-taking in meetings, recalling an initial embarrassing experience with the technology. They express discomfort with the presence of AI note-takers in meetings, particularly when they are not on the call, and mention the ability to remove attendees from the meeting. The conversation also touches on the process of removing participants, which Speaker 1 demonstrates is possible by accessing the participants list.”

    At which point, it was booted from the meeting and sent us this summary.

  11. Phony Genius*

    I think we will soon be seeing malware disguised as AI bots infecting systems. And the malicious AI will train itself to look like the legitimate AI, so you will never be sure what’s safe.

    1. Antilles*

      Is that not what this is?
      The bot was originally sent via an inviting-looking link, which installed some software on one computer, then secretly got access to one of her password protected accounts, then used that access to secretly infect another computer, then the bot sent everyone a link so that it could spread further.
      That feels like a pretty textbook definition of malware right there. The only difference is that as far as we know this AI text-bot hasn’t installed ransomware, downloaded viruses, stolen bank information, stolen passwords, etc.

      1. Your Former Password Resetter*

        Yeah, this sounds like it’s just spyware. Except the spyware is the tool itsself instead of piggybacking on separate link or executable.

  12. Rage*

    This has HIPAA implications as well. I’m sending this article to our compliance manager, because even though some of our people are embracing AI tools (people who are not me, for the record LOL), they think of it as an “opt-in” type of service. They may be totally unaware that such things may propagate via phishing. It wouldn’t even seem all that unusual to be like “oh, how handy, a summary of the meeting. I’ll just download that – oh, gotta make an account first, that’s fine – and be on my way.”

    Everyone read your T&Cs!

    1. AI surprise*

      This is exactly it – I can see them being a potentially useful tool, but this one was more like a phishing bot and totally unwanted. And then it emailed me to participate when I booted it!!!
      So if you see a notetaker log on it’s really appropriate to ask if it’s supposed to be there. And beware opening any links or summaries sent to you by an ai notetaker.

  14. Anon HR*

    We have gotten notification that xyz wants to record the meeting via Zoom for note taking purposes. It says it is an AI assistant that helps teams record and transcribe meetings. You can tell it not to allow the recording.

    In this particular case it’s a bunch of industry HR people talking about challenges, etc. so it’s always a no for us as I’m sure you could imagine.

    1. Rage*

      That might be the case for those that are integrated with Zoom – but others may be able to bypass that entirely.

      Especially since the “wants to record the meeting” occurs at the time when recording starts – and that information is NOT repeated when additional participants log on. By starting the meeting first, bots may be able to bypass that.

      We moved away from Zooms and are using Teams, since it is bundled with our O365 licenses anyway, so I don’t really know how Teams does it as far as notifiacation goes.

      1. RabbitRabbit*

        Yes, my hospital had our computer security department look at all of this. Essentially most of us are now limited to Teams usage for anything internal/hosted by us. That’s outside of some niche cases (and mainly limited to the educational setting) where there are agreements already in place and under oversight.

        Teams has a transcribing option as well but that’s baked into the software and can be toggled on/off. It also warns you about recording being in progress.

  15. The Ginger Ginger*

    I know there are a ton of shady versions out there, and a TON of use cases where they are a problem and/or unwanted, and I am adamantly opposed to almost all of the tech that calls itself AI right now, but there are meetings and implementations where these notetakers (the non-shady ones) are helpful. I sit in a lot of tech and product related internal meetings and it’s been useful to have a notetaker app. We have one that links all the line items back to the timestamp in the meeting recording where it was discussed and summarizes action items, and it’s been super useful when discussing product requirements and meeting on product documentation.

    So yeah, beware shady exploitative unethical ones, and certainly don’t use them or allow them where they’re a privacy concern or unwanted, but the good ones do have a use and can be very helpful.

    1. The Ginger Ginger*

      It’s also nice to not have people expect me (often the only woman in the room) to be the defacto note taker in these meetings. I take my own notes a lot because I have product responsibilities, but it’s helpful for me to have a back up for myself, and for others to not rely solely on me for their own notes.

    2. Nonsense*

      There is no such thing as an ethical open-source AI. And all the available note taker bots are open-source.

      1. The Ginger Ginger*

        Which is why I don’t like basically any of the actual AI stuff out there, but my work is in tech and heavily invested in using AI – as much as I don’t like it, it’s becoming a requirement of my job. I’m trying to find the tools I can to check the Using AI at work box in the least harmful ways possible.

        1. Rage*

          Yeah that’s my take on it entirely. I have stated to my superiors that if any AI (distinct and separate AI; I can’t help what MS and Google do, and they are already on my personal devices anyway) becomes mandatory for my job, I will no longer connect my work laptop to my home internet network. That means I’ll be in the office every day – it’s not a huge deal-breaker for me in that situation – but it also means that I will not be able to complete last-minute projects over the weekends.

    3. AwfulAccuracy*

      In addition to legitimate privacy and security concerns, I have yet to find an AI note taker producing results that bear any resemblance to what was actually said (data/tech/policy/health IT)

      1. wrap me up baby*

        We had one in a meeting recently, The summary indicated that I had agreed to do an action time, when the truth is that I had suggested the action, but it had not been assigned since we will be revisiting things in the new year. I’d be super annoyed to find out people thought I’d agreed to do something due to this sort of shady and not accurate summary.

    1. anonprofit*

      I was wondering if it was Read AI while reading the OP! Looks like it was Otter for them but one of my colleagues uses Read and it sends those same emails telling you to sign up to get the summary.

  17. Sharpie*

    This seems to me to be completely noncompliant with GDPR, so even more underhanded for those of us living in Europe.

  19. CzechMate*

    My husband Fergus’ work uses an AI notetaker tool. He also has a rolling chair on high-pile carpet in his home office (this is important). This was one of the AI transcripts a few months back that he shared with me:

    Introductions and Technical Issues (04:16-09:42)
    -Fergus accidentally fell off his chair during introductions.
    Fergus’ fall of his chair was a humorous moment during the meeting.
    It was mentioned in a light-hearted context to keep the atmosphere relaxed.
    The incident was recorded for the transcript, indicating its significance in the meeting dynamics.

  20. AI? No AIYEEE!*

    It is not at all uncommon where I work to use AI to transcribe/summarize every meeting. It’s considered a good thing/efficient/desirable/the best thing since bread, much less sliced bread. I find it quite helpful sometimes, to be honest, because I’m terrible at taking notes myself (my ADHD brain goes off on tangents and I miss things). But I also work somewhere that is selling AI to clients all over the world, so it’s an assumption that it’s a good thing. If you work in a lot of the corporate world (especially any companies in the $1B+ range), trying to be anti-AI is a good way to find yourself on the outside looking in.

    1. The Ginger Ginger*

      Yeah unfortunately in the tech space it’s becoming more and more of a job requirement. I’m trying to find the least objectionable ways to use it so I can check the box at work without having to use things I find completely unpalatable (which is almost all of it). The meeting note taker is one that works for that as much as anything can.

    2. Zarniwoop*

      A note taker your company chose and controls is great. A note taker from an unknown third party is a security disaster.

      1. Antilles*

        Especially in this case, where the note taking bot did this:
        She didn’t want it and it followed her from her work computer to her personal one. Presumably it got some sort of access to her zoom account, and she didn’t intend any of that!
        A program which installs itself on a completely separate machine without the user knowing about it is a major security concern.

        1. Czhorat*

          It might not literally follow; it’s probably a cloud-based application that gave itself access to her calendar when she signed up, so it’s just inviting itself to all of her scheduled meetings.

          Not that this is any better.

        2. Nah*

          Isn’t this like, the very definition of malware? I suppose it’s not disguising itself as not-an-AI-tool, but worming into your system and injecting itself into other computers through yours is terrifying from a security perspective, especially one you don’t know for sure isn’t *just* harvesting your conversations.

    3. Rotating Username*

      A lot of people don’t have to try to be anti-AI. The wholesale theft of basically the entire learning corpus to train all the large LLMs, not to mention the exponentially increasing power usage they require from fossil fuels, pretty much makes them automatically anti-AI.

      Of course the technology has positives. It can do amazing things. Doesn’t change the fact that most of its marketing is, at best, snake-oil hype, and that on net balance, it is helping destroy the world.

      1. A Girl Named Fred*

        THANK YOU. I was trying to find a way to say that that didn’t sound combative, and you phrased it way better than I could. The more we accept that it’s inevitable as it is, the more it can continue to be the way it is.

      2. I'm just here for the cats!!*

        “the fact that most of its marketing is, at best, snake-oil hype,”

        +1000

      3. Indolent Libertine*

        So much this, particularly where generative AI that produces images is concerned. “Trained” on databases of images scraped from the internet, regardless of copyright, without the consent of the artists, or any compensation to them, or any credit to them (except where you can often see traces of signatures left over from the original images), and with no way to opt your work out of being used this way. Straight-up theft.

    4. StarTrek Nutcase*

      I don’t understand how companies don’t see this as a security risk. Corporate espionage and malicious hacking aren’t uncommon, and it seems like AI note taking just is another possible entry point. Why in 2024 would anyone trust a company or creator of this type of AI to not “share” info? And that’s ignoring someone stealing from the AI source. FFS, at least once a year, I’m notified of a data security breach from a bank, healthcare provider, etc.

      As staff & a human note taker, I’ve had access to lots of highly confidential info discussed in meetings: drug research, patent submissions, attorney work product, patient data, etc. Some of which would have been of great value to competition.

  23. Jennifer Strange*

    I just recorded a video audition for myself singing and when I was rewatching it I noticed on the right an AI bot had been taking notes! I guess good to know my diction was clear enough that it marked them properly???

  24. Czhorat*

    I walked into the same “sign up to see the meeting transcript” trap; two days later on a rare day off I got a call from one of the partners in my company asking me why I had an AI bot attending a meeting on my behalf! I apologized profusely, explained that it was NOT intended, and proceeded to unsubscribe/delete.

    It’s my fault for having given it permissions without thinking, but it’s a very easy mistake to make, and a bit of a predatory behavior on their part.

    1. AI surprise*

      Did it stop showing up after you unsubscribed, or did you have to do something else to scrub it from your meetings?

  25. Zanshin*

    This is scary stuff. I hope you don’t mind that I forwarded this post to the good reporters at Ars Tecnica.

  26. I'm just here for the cats!!*

    Does anyone know where I could go to look more into these bots? I’d like to find out more info to give to my coworkers who work with university students. The school is looking into an AI policy and we’ve been talking about it in our department. I can see students using the AI writing bot in their online classes and not thinking about it when they meet with their counselor.

  27. Anon today*

    Our company has a clear “no AI until security approves the app” policy, even for apps we already use.

    I found out someone was testing out a note taker for our conference platform when I received a transcript of a meeting that appeared to be 20 minutes talking about a current TV show. I hadn’t been attending that meeting but I cohost another meeting with the AI tester so I used to get all kinds of emails that weren’t for me, related to meetings.

    On the up side it did encourage that host to review their settings and remove me from meetings I don’t need to know about!

  28. Roland*

    > The person said DON’T DO IT. It turns out that’s how she got “infected” by the bot. She didn’t want it and it followed her from her work computer to her personal one.

    I don’t doubt they have misleading messaging that encourages consent, but I have to say I doubt it “infected” her and “followed” from one computer to another. That’s just not how it works. She has to have linked it to whatever account she uses to log into zoom. If she uses the same account for work and home that’s why it’s in both places. If she doesn’t, she is probably still logged in somehow with one email on the computer she didn’t mean to be logged in on.

    1. AI surprise*

      I don’t know exactly how it did what it did, but I thought it was relevant that switching computers didn’t matter; being in someone else’s meeting (not one she organized) didn’t matter either. I imagine it somehow got her to unknowingly give permission to it to access something cloud based; I don’t know if it was a calendar with all her meetings or if it was her zoom account or what. But I think the fact that it was easy to unintentionally do that was relevant; further however it was operating, it was also able to email me. So it got access to a lot of info without it being her intention.
      This was a person who is pretty savvy / heavily involved in an online patient support group organization, if it happened to her I imagine it’s easy for it to happen to anyone who isn’t forewarned!

    2. AI surprise*

      … my guess is it got access to her calendar, which I suppose has emails of other participants, and then invites itself to meetings and emails everyone this “summary” that encourages them to log in and then integrates itself into their own calendar and automatically turns itself on unless they carefully opt out or something. Maybe the only way to see the “full summary” is to opt in, IDK, I certainly didn’t log in.

  29. Little Bobby Tables*

    Yikes. If the note taker app is shady enough to install without consent, I wonder if the company providing it is also shady enough to offer the summaries for sale to any interested party

  30. Umiel12*

    I would love to know what happens when the only attendees to show up are the note-taking bots. It might make for a more interesting meeting than most of the ones I attend.

Comments are closed.