> Let's imagine for a second that the whole AI craze doesn't exist, but you still would want to real-time note-taker, what would you do? Indeed, you bring a literal third person to the table. That will just be sitting there, listening in on your conversation and writing everything down.
That's what secretaries were, and this happened in pretty much every significant business meeting for a long long time.
>and this happened in pretty much every significant business meeting
i think the point being made is that you typically didn't bring your secretary with you, notepad and pen in hand, when you went down to the local coffee shop to catch up with someone.
With remote work the equivalent e.g. a quick social chat on zoom or 1-1 meeting is. We should be picky as to which ones we bring a not taker too.
As for a real coffee shop? There is a YC working on a bolt on recorder for your phone so yes. And black mirror is starting to look like a documentary at this point.
i have no idea! i have never encountered it. the original blog comes across as a bit larp-y to me, but on the other hand, some weird stuff happens in some circles and/or cities.
There could be a future, in which, during a date, each person granted their personal ai to take part in the conversation and also to exchange information among them, probably first to establish trustworthiness of the person in front of you.
So… it’s essentially computers entering into relationships on behalf of their human subscribers thus acting as a personality shim between each one of the people ensuring compatibility which will abruptly expire upon termination of the subscription…
This article and the one it is responding to are explicitly talking about non-business meeting contexts. They are talking about personal conversations, medical appointments and much more casual business contexts like informal interviews. Places where there would never have been a secretary present.
medical appointments is a perfect use case for ai note takers. Imagine you have an elderly parent who lives hundreds of miles away. You can't always be present for their medical appointments, yet, the outcomes of those appointments are critical, and sometimes information is shared in a way that the elderly parent cannot be a trustworthy source for that info. What do you do?
Record the audio, transcribe it later. Your AI note taker will mangle some of the details in a subtle but wrong way. Or not, who knows? But if you have ever transcribed anything with whisper, it is both amazing how well it works, and how it can go completely off the rails if there is a prolonged pause or noise. It will just make up words, fill the gaps. I am assuming the transcription process for an AI note taker will not use anything fundamentally different than Whisper
You make them get the results and guidance in writing, and if they forgot to get it in writing, you have them contact the provider and get it in writing. You don't send a tape recorder.
To a physio therapist (the article example) that is a business meeting. That's what's strange here. That person isn't there because their you're friend: they're getting paid to be there, which makes it a business meeting, which makes a secretary, human or AI, appropriate.
Does /anyone/ take notes in a personal context? I don't take notes when I catch up with friends, but neither does anyone else. It's a complete non-issue.
Secretaries weren't invited to one on ones and interviews generally. This was something for only the highest places employees and not say something an IC would be used to.
The difference being they were directly paid by the people involved and their notes weren't whisked off to outside companies with very loose privacy policies around your inputs.
Secretaries are compelled by the same privacy and disclosure laws as doctors - AI notetaking apps may or may not be (it'll take examination of the product on a case by case basis) and the public is, at this point, used to AI companies blatantly lying about privacy, confidentiality, training sources, reuse of conversations and pricing - the good faith is gone as a default and in a setting like a HCP that is a terrible place to start from.
Anyone handling PII needs to be trained in HIPAA compliance and authorized in the same manner as the doctor themselves. There is no explicitly separate HIPAA like law though if your handling of the information is limited to a particular subset of the information (e.g. knowing that a patient has an appointment with a renal specialist and knowing their address information but not knowing the topic of the appointment) then there may be a relaxation of the requirements - but patients are also blabby so usually they're treated like anyone else.
When it comes to AI note-taking or other AI services that interact with PII the exact same requirements and data handling standards need to be adhered to as if the tool was an employee with proper certified training (or, in the case of a tool, review and certification) and appropriate contracts and disclosures.
I can't really provide a more specific answer because the topic is so broad and while I'm familiar with the process from one end I am not an expert on the process in general.
In personal, light-hearted or even most 1 to 1 meetings I'd agree.
But my note taker has been an incredibly helpful piece of technology in professional meetings where domain decisions had to be recorded, high-throughput arguments were exchanged, ideas brought up and todos just verbally agreed on.
I always ask people, kindly and honestly with a lot of room for denial, for agreement to record meetings with them. No, I don't do it in meetings where I know personal topics could come up or be relevant. But those are more on the rare side. In most of my professional exchanges it hasn't posed a problem for anyone.
Let's try the suggested advice and ask why five times.
> Why do you even need an AI assistant here?
To take notes on what we talked about.
> Why do you want to do that?
Because I want to retain the contents of this conversation, and I don't want to be distracted by note-taking. I want to be in the moment and also have a record of what we talked about for later.
> Why do you want a record?
Because I expect that what you're going to say is valuable enough to want to reference later. Perhaps you will give me the name of a cool podcast, or you'll give me very good, detailed advice. Perhaps you'll mention when your upcoming birthday is or a favorite brand of a product, and it'll be useful for gifting.
I use it in my professional appointments. If someone asked me not to, I'd turn it off. It makes the flow of the appointment much better than if I try to write the notes as we go, and if I don't write then and there, something will probably be left out.
Edited to add, I don't see what problem the OP has, unless they're just uncomfortable saying no.
These AI Note-Takers can also mangle the summaries. A few weeks ago I read anecdotes here about a doctor getting completely wrong information about a patient, and a manager getting upset because he was depending on a summary of something a client never said or agreed to, but the AI summary said he did. These things are downright dangerous.
Now come the replies saying, "as if human note-takers never made mistakes!"
Moreover, they capture all the smalltalk perfectly, but tend to plausibly mishear the parts that matter most - terminology, abbreviations, names of people. There's no mystery why this happens, of course, but the only transcript I trust is a transcript I personally checked by listening to the audio and fixing mistakes (doable at 1.75x speed, so not that bad). I catch outrageous mistakes sometimes, mistakes that completely change the meaning of what's being said (up to capturing the opposite of what's being said). So, all in all, even though modern speech2text models are really impressive, I am not sure the utility of _completely_ automated transcribers outweighs their dangers today.
It would be nice if they would at least say __unintelligable__ when they couldn't get things, but alas. I also notice it with YouTube auto generated subtitles and translations. They misrepresent SO much.
Actually I think one of the main values of note taking isn't that you get notes, it is that engaging your brain to summarize points makes you understand and remember the points better, even if you throw your notes away right after.
It is like writing: One of the main purposes of writing (for me and many others) is that it forces you to structure your thoughts. You could give that task to AI, get the resulting text, but your loss is the lack of clarity you could have gained from doing it yourself.
While I understand the dislike of the concept, I find myself liking my own universal notes of meetings I attend on a daily basis: It's a democratization of the concept of secretaries.
While not all meetings need action items and summaries, having high level notes or the full transcript available when you actually get to implementing a task, or having an agent go through the dozen meetings you had on a project to identify its evolution has made finding and keeping knowledge up-to-date much better in my own experience.
It helps account for bias towards the most recent candidate in interviewing as you can now compare all notes, helps recall some niche details that were said in passing for projects, and holds everyone accountable to what they said.
I still haven't found a perfect solution to actually storing and querying this knowledge, single meetings are easy but wiki-fying them is my end goal, but I'm very happy about the direction dictation, transcriptions and AI parsing of all this knowledge is going.
I have usually permitted that when it comes up and sometimes declined it. But when I permitted it I didn't have to worry about them entering everything I say into a permanent history from which a random sentence could get used, datamined, hallucinated into something that could be used against me years later.
I have mixed feelings on this (I have some relatives in the medical industry). On the one hand, having a professional scribe is absolutely a positive for the provider, provided they have been trained and are accustomed to that provider. They take away significant cognitive load from them, leaving them able to focus on the patient.
With the proliferation of AI note taking, this advantage purports to have been democratized, but I'm not quite convinced. Since AI summaries are far from infallible, a mistake is bound to sneak in here and there (note that these aren't mere transcripts, but summaries split into sections). The provider may or may not go in and clean up your AI notes afterwards, any mistakes made by AI are effectively disowned in terms of responsibility, and admin will still pressure providers using these note takers as leverage to be able to see more patients than otherwise possible (admins want to see both lower costs and higher patients seen per day).
When you refuse this type of service, you're demanding a higher bar for your notes, but it comes at the cost of a distracted provider (who has grown accustomed to AI note taking, and only has so many hands, so they have to go back to the keyboard every now and then after checking your body).
In summary I think it comes down to how much you care about note quality versus care quality, which is likely different per person. I don't have any allergies, am not on any medications, and generally only go in for routine checkups, so in my case the notes are more or less a bureaucratic requirement that I'm happy to do away with cheaply. For others this may not be the case, and having quality notes may be critical to their care, in which case they should definitely refuse AI scribes.
Of course, none of what I said goes into privacy, which is a significant matter. However, "iPad scribes" which are essentially third party contractors remotely taking care of the notes already exist, so those concerns which arise even without the use of AI are a bit of a different topic.
AI only needs to be better than a human, which is also fallible.
But I would only accept AI note taking if it was reconciled (in a timely manner) by a human listening to the audio recording and verifying accuracy. Without that, it's an unacceptable loss of accountability. Even with that, privacy concerns are abound with frontier models.
STT local models offer different concerns, but generally are pretty good at transcribing in a quite environment.
Capitalistic pressures reduce quality of health care far too often. Everything from artificially limiting how many doctors get degrees, to pushing doctors to see more people per day from such tech. The enshittification will continue until moral improves.
I agree with some points being made and disagree with others.
Professionals should be using the best tools they're comfortable using. If that means your therapist can ask better questions and make better use of their time because they have a STT model running, I think that's great, counter to the author. Importantly though, in therapy especially, your comfort matters more than time efficiency, and can understand asking for it to be disabled.
Catching up at a coffee shop does not need STT, I agree with OP...but are people really doing this? Sure, there's a few AI obsessed geeks doing it I'm sure, but is that really a main use-case of these devices?
This is a reply to someone else's blog post (on a different site - odd concept, I kinda like it though). It may add a lot of context to read the original blog post https://www.joanwestenberg.com/p/im-begging-you-to-leave-you... first, which is linked in the header but easy to miss, and then read this reply.
I rarely have online meetings, but what annoyed me most last time I did was that I had multiple AI note taking services magnanimously send me multiple emails each telling me how if I signed up for their service they could share their meeting notes with me. It wasn't even that important of a meeting!
Well, had a meeting with a VC that literally brought his note taking 3rd person to the meeting, did not even say hello. Tbh, I still find it creepier than an AI transcriber.
did you use AI for your site because it ain't loading... so based on the title... I have no problem with people using them to help their memory, and if they kept it local, but I have a problem with monopolistic and exploitative companies with private data, that likely ends up in a govt db.
Knowing both doctors and nurses before AI scribes, actually hand writing notes isn’t usually done in front of the patient. They make quick notes so they keep up with the conversation and then when they need to do long form, they reference their scribbles but do the rest from memory at the END of the day.
For every one person that says they mind, it adds likely an additional margin of error AND time to their day. So I really disagree with making people’s lives harder. That’s why technology exists…
And more and more of these systems are on device models, not cloud compute. So I don’t see that argument either
I don't think the issue AI note-taking per se. It is whether a conversation can still be candid and confidential once everyone knows it may become permanent, searchable, and shareable. Big tech companies don't exactly have a clean record on privacy.
Just saying no sounds simple enough. But social pressure makes that very difficult.
The normalisation of constant recording really has me worried, everything is so monitored now. Privacy just keep get more and more eroded for the sake of minor convenient features and products. At what point do we start deciding these trade off are no longer worth it? And is it even too late to make that decision?
I think most people here miss the point. IF the AI Note taker would be only on device, it would be less creepy. But at moment everything goes to some companys (and so also more or less direct, to the government too).
I can understand, especially in a medical context, being bothered by AI notetaking specifically because it implies your private information being handed off to a third-party where you cannot control it, rather than that an AI tool is processing it. That said, it is common practice in business to take notes or record calls. I take notes either by hand or type-written in nearly every meeting that I'm in. I often record my meetings. When it's entirely within my company, we have an internal (e.g. not third-party) AI transcription tool that I enable. When meeting with external entities, I either write notes, record audio, or have a designated person with me to write notes.
Taking minutes/notes in a meeting and then being accountable to follow-up on any action items out of that meeting is just standard business process across industries and across the globe. That's not what the author is referring to with their therapist, that's a very different context, but in the context most people on HN find themselves in, having someone take notes is not only not a big deal, it /should/ happen.
> How fucking creepy is that?! How fucking awkward is that? And before you start talking with whoever you're catching up with, you ask "You don't mind, do you?"
In living memory, I had a HUMAN notetaker in important meetings. After secretaries left the world went to hell and topic experts and engineers were expected to have social graces, everything got worse. We invented new religions like agile to make up for a good old organized secretary.
So - no I won't apologize. My memory is that of a catfish. I see a moving object and i head towards it. Note takers are invaluable, human or not. And AI or NOT voice to text is NOT new.
Lastly, it's worked out to keep everyone honest. I work with clients, we have calls, they're long. I just had a client pull ME up in an old recording agreeing to do something after I said 'no thats out of scope'. So its nice to see some accountability.
This particular article concerns an AI notetaker being used in the context of healthcare - there is a well founded reason that health information is so stringently controlled which is to enable patient comfort at sharing potentially embarrassing or uncomfortable details. If you're receiving PT there is a large potential to feel shame over an inability to do things that were previously trivial[1].
Once upon a time the notes may have been recorded by a staff member of the doctor's office or by the doctor themselves (usually after the meeting). Budgetary constraints push HCPs towards cutting staff and those rolls are being replaced by AI and that is not okay.
Staff, Nurses, Doctors are all under the clear guidance of HIPAA and understand their responsibility towards patient privacy - it isn't a perfect system and there are notable disclosures and violations that have happened in the past but once third party systems are involved - especially non-deterministic third party systems - then the client's understanding of privacy may be severely compromised.
I love voice to speech and meeting summarization for thinking sessions with coworkers where maybe someone is motivated to take notes (and better participates through that action) but the emphasis is on everyone being present and being able to talk freely. The doctor's office is a fundamentally different environment, though.
1. Aside - an unnecessary shame - no one should feel guilt over trying to overcome a disability.
You absolutely could build such a tool and I have seen HIPAA authorized model sandboxes that do definitely exist. However, most off the shelf note-takers would not so qualify and it's a minefield to build one with out of domain third party resources.
I think it's fair to say the non-technical public has been repeatedly burned by AI companies selling something different from what they're advertising and it's fair to be skeptical as a layperson about your ability to tell a properly HIPAA compliant tool from a non-HIPAA compliant tool.
It has occurred to me over and over for the last several years that many of the senior engineers at my company would be substantially more productive with some sort of assistance from someone who specializes in having executive function skills. One such person given four or five engineers to manage could do wonders. And they'd also be the best source of feedback for hard-to-measure performance evaluation information like "is this senior engineer actually working on anything most days?" But executive assistants are a privilege reserved for only the people who are at level X and up.
somehow my paltry and boring CRUD business specifics seem a lot less important than recording a doctors recommendations as to how to treat my wounds.
I have a son and we have to sometimes split appointments across me and my wife. I wish i had a recording or automated AI notes that came to my email (no not to a captive portal I can lose access to) so i can read what my doctor said about my son without it getting filtered through my wifes memory.
But the quoted example and your counterexample are very different cases! It can be creepy to use an AI note taker when "catching up" and totally a good idea for hour-long client calls.
I strongly disagree that agile was invented to make up for the absence of secretaries. Agile was invented to make up for the absence of omniscience (and absence of the recognition of non-omniscience). Secretaries weren't going to make that work.
That's what secretaries were, and this happened in pretty much every significant business meeting for a long long time.
reply