Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
-
@Em0nM4stodon What makes this doubly shitty is that we had ML to do transcription for accessibility purposes (which usually did an ok job for some languages), and now most of these meeting apps instead just shove the lying slop machine in there for shittier translation, garbage summaries, and privacy violations.
I complained at work and was told it can't be turned off because we need transcripts for accessibility. Fuck.
@sharpcheddargoblin @Em0nM4stodon Legitimate transcription for accessibility requires a local speech to text engine with no persistent recording of the input. Which already existed before "AI". Trying to use this as a lever to get "AI" in the door is entirely bad-faith.
-
@dalias but the very technology you’re lauding is built using the very same techniques as the dreaded “AI”. To bring up AI is bad, when the problem you really have is persistence is bad faith.
Even just talking about “AI” as if it is anything other than vapid marketing speech not only does the billionaire work for them, it clouds your own argument and renders it risible
-
@dalias but the very technology you’re lauding is built using the very same techniques as the dreaded “AI”. To bring up AI is bad, when the problem you really have is persistence is bad faith.
Even just talking about “AI” as if it is anything other than vapid marketing speech not only does the billionaire work for them, it clouds your own argument and renders it risible
@jonathankoren No it is not.
The only "technology" "the dreaded AI" is built upon is the technology of the cold reader, the mentalist, the con man, the "technology" of understanding how to make gullible humans believe you're doing something you're not.
And that's what you're trying to do now, and it's not working.
-
@jonathankoren No it is not.
The only "technology" "the dreaded AI" is built upon is the technology of the cold reader, the mentalist, the con man, the "technology" of understanding how to make gullible humans believe you're doing something you're not.
And that's what you're trying to do now, and it's not working.
@jonathankoren Oh, wow, I love Mastodon profile notes. Already had on this one:
"Stan for AI shit, willfully misinterpreting words against it"
Time for a block.
-
@mariyadelano @sharpcheddargoblin @Em0nM4stodon Keyword to search for is probably "local model" or "offline". It's not something I've needed so I haven't looked into what's currently available.
-
Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
In certain cases, it can even breach two-party consent laws, and be a criminal act.
Ideally, never use these privacy-invasive tools ever. But if you do use them, you must ask for everyone's consent first, and respect their choice fully if they refuse to be recorded.
@Em0nM4stodon I attended a gathering that was cast online and saw a notice about an AI bot. I have since stopped attending any gatherings of that group.
-
@mariyadelano @sharpcheddargoblin @Em0nM4stodon Keyword to search for is probably "local model" or "offline". It's not something I've needed so I haven't looked into what's currently available.
@mariyadelano @sharpcheddargoblin @Em0nM4stodon The Common Voice project https://commonvoice.mozilla.org is one of the few open, ethically sourced data sets I'm aware of, but I'm not sure if there's a transcription engine built on it.
-
@realcaseyrollins @jackryder@infosec.exchange @Em0nM4stodon You mean the refusal string itself? It was in MisAnthropic's Claude docs.
-
Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
In certain cases, it can even breach two-party consent laws, and be a criminal act.
Ideally, never use these privacy-invasive tools ever. But if you do use them, you must ask for everyone's consent first, and respect their choice fully if they refuse to be recorded.
@Em0nM4stodon@infosec.exchange wonder if 'by continuing you consent' messages that get automatically sent comply with this
-
Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
In certain cases, it can even breach two-party consent laws, and be a criminal act.
Ideally, never use these privacy-invasive tools ever. But if you do use them, you must ask for everyone's consent first, and respect their choice fully if they refuse to be recorded.
@Em0nM4stodon Haha the interview I had with someone where she said "Do you mind if I record? It won't go anywhere, it's just to help me transcribe."
And I said "Oh sure, that's interesting, will you transcribe them yourself?"
And she said "Oh I'll get the AI to do it."
I *really* should have pointed out that in that case it will in fact go everywhere, but I was suddenly beset by a soul-crushing nihilistic ennui and just went with the rest of the interview. Sigh.
-
@jonathankoren Oh, wow, I love Mastodon profile notes. Already had on this one:
"Stan for AI shit, willfully misinterpreting words against it"
Time for a block.
@dalias @jonathankoren@sfba.social I have jonathankoren blocked also.
I wish I took notes now. -
Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
In certain cases, it can even breach two-party consent laws, and be a criminal act.
Ideally, never use these privacy-invasive tools ever. But if you do use them, you must ask for everyone's consent first, and respect their choice fully if they refuse to be recorded.
@Em0nM4stodon Well said.
-
Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
In certain cases, it can even breach two-party consent laws, and be a criminal act.
Ideally, never use these privacy-invasive tools ever. But if you do use them, you must ask for everyone's consent first, and respect their choice fully if they refuse to be recorded.
@Em0nM4stodon Does the privacy invasion start there? Or does it start when people decide to record the meeting? Or does it even start when someone forces you to use these tools?
I'm asking because I'm actually not sure. Does the usage of AI tools really make matters worse than it already was before? Or do the same privacy issues already apply without them?
-
(Not towards you) But generally I mean I don't know why people need AI involved, just ask to take a video/audio of the meeting.
@YourShadowDani @Em0nM4stodon If the audio/video of the meeting gets recorded (maybe visible to you and usable by you, or completely hidden), hasn't the privacy invasion then already happened? Getting a transcript from some AI doesn't seem like an additional problem, assuming you were already okay with the former. You just get at least something back, you could say. Or not?
(I'm not defending all that btw)
-
Just so you know, using an AI notetaker during a meeting without priorly asking for consent from all participants before starting it breaches many if not most privacy laws around the world.
In certain cases, it can even breach two-party consent laws, and be a criminal act.
Ideally, never use these privacy-invasive tools ever. But if you do use them, you must ask for everyone's consent first, and respect their choice fully if they refuse to be recorded.
@Em0nM4stodon is that true for any notetaking software or voice and video recording? How is AI any different?
-
@Em0nM4stodon What makes this doubly shitty is that we had ML to do transcription for accessibility purposes (which usually did an ok job for some languages), and now most of these meeting apps instead just shove the lying slop machine in there for shittier translation, garbage summaries, and privacy violations.
I complained at work and was told it can't be turned off because we need transcripts for accessibility. Fuck.
@sharpcheddargoblin @Em0nM4stodon "lying slop machine" is just "algorithm aversion". Software can make mistakes, for gods sake autocorrect or even "glide recognition" screws up regularly, or rather has less than 100% accuracy. We don't condemn all of its efforts because of that.
-
@jonathankoren No it is not.
The only "technology" "the dreaded AI" is built upon is the technology of the cold reader, the mentalist, the con man, the "technology" of understanding how to make gullible humans believe you're doing something you're not.
And that's what you're trying to do now, and it's not working.
@dalias @jonathankoren ffs get real
-
@dalias @jonathankoren@sfba.social I have jonathankoren blocked also.
I wish I took notes now. @anolandria @dalias you can use AI for those notes, apparently
-
@sharpcheddargoblin @Em0nM4stodon Complete ill-informed bollocks. A case of algorithm aversion should not stop anyone with intelligence from discovering how things work.
-
@anolandria @dalias you can use AI for those notes, apparently
