This type of software only comes to use in those cases of journalists/writers who have a very good OP SEC, aka they can't be traced or spied upon. Hence arises the problem of identifying them in other ways.
But, generally the CIA/NSA etc, don't need that type of software for 99.9% of people. They aready know your beliefs, convictions, what you like, what makes you angry, what addictions you have, your fetishes, if you can be corrupted by power, money or status etc
And I'm sure they have other software to create propaganda geared at certain sections of a population in order to divert their attention, manipulate them emotionally etc. These tools are used in these so called modern revolutions, see ARAB SPRING, where they use social media algorithms, to manipulate and feed certain true or false info to the users and incite in them a certain needed emotional state that would make them more helpful, or indifferent to CIA agents overthowing their government.
You can already copy and paste a paragraph of something and you will easily find the source. AI/software can do this automatically, with no effort. And can do this for all paragraphs, or sentences in an article. Furthermore, linguists, or even experienced writters and editors, who are well read and study language deeply, know and are able to discriminate different styles of writtings. AI/software can do this better, and faster. So, searching for a match among say 10million ppl (overestimate of English speaking high profile, courageous informed journalists/activists in the entire world) is a very very easy task.
Btw, we are not talking about the average person, much like yourself, who writes a logically deficient sentence on some obscure forum.
There's a yes and a no there. Let's say one wants to track only specific targets like journalists, or people inside organizations who could become leakers. What you do is gather things those specific people wrote, then train the AI to look for similarity in things on the web. It then is a means to find things to keep an eye on.
There is already various software that can analyze linguistic signatures. I have little doubt this is already deployed by certain agencies.
This type of software only comes to use in those cases of journalists/writers who have a very good OP SEC, aka they can't be traced or spied upon. Hence arises the problem of identifying them in other ways.
But, generally the CIA/NSA etc, don't need that type of software for 99.9% of people. They aready know your beliefs, convictions, what you like, what makes you angry, what addictions you have, your fetishes, if you can be corrupted by power, money or status etc
And I'm sure they have other software to create propaganda geared at certain sections of a population in order to divert their attention, manipulate them emotionally etc. These tools are used in these so called modern revolutions, see ARAB SPRING, where they use social media algorithms, to manipulate and feed certain true or false info to the users and incite in them a certain needed emotional state that would make them more helpful, or indifferent to CIA agents overthowing their government.
Considering how many people there in the world, I am sure the accuracy is piss-poor
You have trash logic. Here's why!
You can already copy and paste a paragraph of something and you will easily find the source. AI/software can do this automatically, with no effort. And can do this for all paragraphs, or sentences in an article. Furthermore, linguists, or even experienced writters and editors, who are well read and study language deeply, know and are able to discriminate different styles of writtings. AI/software can do this better, and faster. So, searching for a match among say 10million ppl (overestimate of English speaking high profile, courageous informed journalists/activists in the entire world) is a very very easy task.
Btw, we are not talking about the average person, much like yourself, who writes a logically deficient sentence on some obscure forum.
There's a yes and a no there. Let's say one wants to track only specific targets like journalists, or people inside organizations who could become leakers. What you do is gather things those specific people wrote, then train the AI to look for similarity in things on the web. It then is a means to find things to keep an eye on.