I’m referring to two friends right next to each other, talking silly shit in person, like what’s the random phone microphone and associated AI think of the words it hears?
We (me and my roommate) just got done joking about a crazy woman that wanted to have 6 babies with me and wanted to move to Canada 😂🤣
I’m not asking about any AI chatbot, I’m asking what do the random phone microphones and associated AI systems think of random jokes it heard?
OP is responding to your original question. If you’re asking if it safe to do it, then you are worried about the consequences of doing it and wonder if you should stop. That would mean that you are “obeying a rule before it exists”.
As to your current comment, you should understand that current “AI” solutions do not think anything. If the current solutions were used for mass surveillance, then they would be used to classify your actions in some predefined way and whomever is using them will use this classification in some (possibly nefarious) unknown way
Yo, you both make good points. Sorry I didn’t explain the joke/banter we were discussing.
We (in person) were joking for around 10 minutes or so about a crazy woman that literally wanted to have 6 children with me and move to Canada.
She’s crazy and was recently evicted, but we continue to joke about it. But it really makes me wonder, what if the microphones and AI shit around hear our silly banter and take it seriously, like thinking I’m about to move to Canada (I’m not).
Long question short, does AI know the difference between serious words or jokes?
i talked about subject like this with someone who is an anarchist and attends kind of protests that can get you in trouble and he wasnt that worried about it. It was in finland though, so there is that too. if i lived in america i would be way more worried too considering what is happening there.
But lets assume that your microphone is recording everything you say to some llm, and everyone else too. imagine the shitshow it must be to sort all that data, considering how unreliable llm are. There is no way to know for sure what the original meaning and intention of words is without actual human confirming it. there would be so much constantly going on that even with decent automation it would be nightmare to manage, i think.
and the subject you joked about is so innocent too. If you were talking about violently opposing the regime as a joke or something else like that, then it would be prudent to practice better security.
I have occasionally tried to investigate if my phone is listening even though it doesnt indicate so, but i havent noticed anything that would point to that. I dont know THAT much about it either though. Though if you have all the bloatware installed and llm stuff enabled then it probably does listen constantly, but i dont think you would have them since you are worried about this.
There is also this nice program called tracercontrol, it lets you know what spying libraries programs have on your phone and you can also cut them off internet with it, and also selectively block the traffic so only necessary traffic is allowed (it does require some getting used to, but its not that difficult imo.) There is some version of it in googles malwareshop called google play, but i downloaded it from f-droid.
There is also similar program called re-think, which is even more heavyduty, though i think it uses battery faster. but at least you can control even better what goes in and out of your phone with it.
And why do you even care what the companies think about you? also, isnt it better less accurate information they have about you? let them think you are moving to canada with 15 kids, to decorate garages with ornamental potatoes and live off the good wibes of the universe.
But lets assume that your microphone is recording everything you say to some llm, and everyone else too. imagine the shitshow it must be to sort all that data, considering how unreliable llm are. There is no way to know for sure what the original meaning and intention of words is without actual human confirming it. there would be so much constantly going on that even with decent automation it would be nightmare to manage, i think.
filtering for interesting words in a transcription would go a long way. I’m not convinced they would need an LLM for this. keyword based targeted advertising was a thing for quite a fewyears before LLMs became widespread.
I have occasionally tried to investigate if my phone is listening even though it doesnt indicate so, but i havent noticed anything that would point to that.
I have friends of all ages who are not really technically adept, and not caring much about privacy either, but who tell me from time to time that it’s like facebook is listening, because of some ad they saw on there. I still don’t have any clue how are they pulling that off, and I hear it too much to accept it being a coincidence.
but I don’t use facebook, I refuse to use their apps, my phone is clean from google too, while theirs is littered with every garbage in addition to the factory bloat, so I can’t do much to figure it out
I’m really confused here, what do you mean?
I’m referring to two friends right next to each other, talking silly shit in person, like what’s the random phone microphone and associated AI think of the words it hears?
We (me and my roommate) just got done joking about a crazy woman that wanted to have 6 babies with me and wanted to move to Canada 😂🤣
I’m not asking about any AI chatbot, I’m asking what do the random phone microphones and associated AI systems think of random jokes it heard?
It doesn’t think, it can only find correlations. There isn’t a person in there forming opinions of its own.
OP is responding to your original question. If you’re asking if it safe to do it, then you are worried about the consequences of doing it and wonder if you should stop. That would mean that you are “obeying a rule before it exists”.
As to your current comment, you should understand that current “AI” solutions do not think anything. If the current solutions were used for mass surveillance, then they would be used to classify your actions in some predefined way and whomever is using them will use this classification in some (possibly nefarious) unknown way
Yo, you both make good points. Sorry I didn’t explain the joke/banter we were discussing.
We (in person) were joking for around 10 minutes or so about a crazy woman that literally wanted to have 6 children with me and move to Canada.
She’s crazy and was recently evicted, but we continue to joke about it. But it really makes me wonder, what if the microphones and AI shit around hear our silly banter and take it seriously, like thinking I’m about to move to Canada (I’m not).
Long question short, does AI know the difference between serious words or jokes?
i talked about subject like this with someone who is an anarchist and attends kind of protests that can get you in trouble and he wasnt that worried about it. It was in finland though, so there is that too. if i lived in america i would be way more worried too considering what is happening there.
But lets assume that your microphone is recording everything you say to some llm, and everyone else too. imagine the shitshow it must be to sort all that data, considering how unreliable llm are. There is no way to know for sure what the original meaning and intention of words is without actual human confirming it. there would be so much constantly going on that even with decent automation it would be nightmare to manage, i think.
and the subject you joked about is so innocent too. If you were talking about violently opposing the regime as a joke or something else like that, then it would be prudent to practice better security. I have occasionally tried to investigate if my phone is listening even though it doesnt indicate so, but i havent noticed anything that would point to that. I dont know THAT much about it either though. Though if you have all the bloatware installed and llm stuff enabled then it probably does listen constantly, but i dont think you would have them since you are worried about this.
There is also this nice program called tracercontrol, it lets you know what spying libraries programs have on your phone and you can also cut them off internet with it, and also selectively block the traffic so only necessary traffic is allowed (it does require some getting used to, but its not that difficult imo.) There is some version of it in googles malwareshop called google play, but i downloaded it from f-droid. There is also similar program called re-think, which is even more heavyduty, though i think it uses battery faster. but at least you can control even better what goes in and out of your phone with it.
And why do you even care what the companies think about you? also, isnt it better less accurate information they have about you? let them think you are moving to canada with 15 kids, to decorate garages with ornamental potatoes and live off the good wibes of the universe.
filtering for interesting words in a transcription would go a long way. I’m not convinced they would need an LLM for this. keyword based targeted advertising was a thing for quite a fewyears before LLMs became widespread.
I have friends of all ages who are not really technically adept, and not caring much about privacy either, but who tell me from time to time that it’s like facebook is listening, because of some ad they saw on there. I still don’t have any clue how are they pulling that off, and I hear it too much to accept it being a coincidence.
but I don’t use facebook, I refuse to use their apps, my phone is clean from google too, while theirs is littered with every garbage in addition to the factory bloat, so I can’t do much to figure it out
To answer your question MUCH more concisely: with a probability, within a context limit.