The Pentagon Wants To Use AI To Create Deepfake Internet Users (theintercept.com) 43
schwit1 writes: The Department of Defense wants technology so it can fabricate online personas that are indistinguishable from real people.
The United States' secretive Special Operations Command is looking for companies to help create deepfake internet users so convincing that neither humans nor computers will be able to detect they are fake, according to a procurement document reviewed by The Intercept.
The plan, mentioned in a new 76-page wish list by the Department of Defense's Joint Special Operations Command, or JSOC, outlines advanced technologies desired for country's most elite, clandestine military efforts. "Special Operations Forces (SOF) are interested in technologies that can generate convincing online personas for use on social media platforms, social networking sites, and other online content," the entry reads.
The United States' secretive Special Operations Command is looking for companies to help create deepfake internet users so convincing that neither humans nor computers will be able to detect they are fake, according to a procurement document reviewed by The Intercept.
The plan, mentioned in a new 76-page wish list by the Department of Defense's Joint Special Operations Command, or JSOC, outlines advanced technologies desired for country's most elite, clandestine military efforts. "Special Operations Forces (SOF) are interested in technologies that can generate convincing online personas for use on social media platforms, social networking sites, and other online content," the entry reads.
Yes, please (Score:5, Funny)
Please create a lot of AIs to populate social media platforms. If we get enough of them talking to each other, then all of the actual people can leave them, and we might have a cultural renaissance as we are freed of our societal addiction to social media.
(...including Slashdot...)
One does not simply social on Slashdot. (Score:1)
Re: (Score:2)
Honestly, being beneath notice for propaganda campaigns is one of the nicest parts of "old-style" internet posting.
Someone might be getting paid to astoturf on slashdot, but if so, they're not bringing the bot farms to the table.
Re: (Score:2)
Re: (Score:1)
The fact that the editors don't strip bad mods of their mod points is sufficient evidence to prove the point that no one cares about this site. Including the management.
Re: (Score:2)
Please create a lot of AIs to populate social media platforms. If we get enough of them talking to each other, then all of the actual people can leave them ...
Make all the AI bots "moms" and "dads", that'll get all the kids out. :-)
It must be said (Score:3)
Stupids, don't announce yr evil plans (Score:2)
...you keep them deep underground where you can cackle maniacally yet nobody above can hear.
They need a refresher course in World Domination 101.
Re: (Score:2)
...you keep them deep underground where you can cackle maniacally yet nobody above can hear.
They need a refresher course in World Domination 101.
What you have to ask yourself is this: If they're publicly popping this absolutely diabolically evil shit? What is it they're still hiding from us? Someone considers this the surface level "let them know, it'll distract them" move. What are they really planning on using AI for? I mean, aside from outright weapons.
With the way our media works now, having social media posts be the drivers of most news, this is fixin' to be an absolute shit-show in the coming years. Maybe we'll finally gut the whole media mach
Re: (Score:2)
They are volunteering to provide the test data for https://it.slashdot.org/story/24/10/16/217207/startup-can-identify-deepfake-video-in-real-time [slashdot.org] and I can't see that as a particularly "bad thing". At the very least we get to see if "Reality Defender" actually does what it claims.
Of course the real test data is on Twitter, let's see if the Muskbot is classed as Real, AI or AU (aka AS)*.
* Unintelligence or Stupidity.
Re: (Score:2)
They are volunteering to provide the test data for https://it.slashdot.org/story/24/10/16/217207/startup-can-identify-deepfake-video-in-real-time [slashdot.org] and I can't see that as a particularly "bad thing". At the very least we get to see if "Reality Defender" actually does what it claims.
Of course the real test data is on Twitter, let's see if the Muskbot is classed as Real, AI or AU (aka AS)*. * Unintelligence or Stupidity.
Yeah, about them providing test data? That puts a nice warm fuzzy on their propaganda creation tool, but it's still a propaganda creation tool.
Re: (Score:2)
If they're publicly popping this absolutely diabolically evil shit?
I hardly think some bots on social media is diabolical or evil, actually its not really even different than having the humans doing astroturfing that is probably already happening.
sit down have a cup of tea and drink it slowly be careful it may be hot.
All this is really going to do, as the pentagon does it as well as every other intel agency the world over is turn up the BS a little bit. As it is if you are taking anyone remotely controversial seriously from people you don't IRL know you are probably being
Re: (Score:2)
If they're publicly popping this absolutely diabolically evil shit?
I hardly think some bots on social media is diabolical or evil, actually its not really even different than having the humans doing astroturfing that is probably already happening.
sit down have a cup of tea and drink it slowly be careful it may be hot.
All this is really going to do, as the pentagon does it as well as every other intel agency the world over is turn up the BS a little bit. As it is if you are taking anyone remotely controversial seriously from people you don't IRL know you are probably being manipulated. Especially if you are making assumptions about how popular something or some idea is or isnt based on count s of likes / retweets / mentions / etc. Rest assured if its important someone has bot or mechanical Turk farm working hard to make it appear they are on the right side of whatever..
Adding a few more and better bots will do nothing but make the already terrible SNR a bit worse. Maybe it make it bad enough it will go the way of CB radio in the 90s and people will just give up. Which is probably what's been happening since the advent of the printing press, I am sure there was a time when simply because someone typeset a message it was afforded some credence, not when guy hands you their 'newsletter' while you are walking in the park into the next waste basket it goes.. Frankly nothing of value will be lost. Web 2.0 has been a net negative for society. For every "useful form on subject" out there is a 1000 flame wars and 10 pages of algorithmic selected garbage to sort thru. It enables and re-enforces everyone worst behavior and social maladies. Some new media or new iteration of something else will become assessment for a time, and then the propaganda people will colonize that, on and on it repeats.
Of all thing things that happen in side the pentagon that should keep you up at night, this probably isnt it.
Based on the cascading avalanche of bullshit already on social media, it'll either do nothing but ripples, or be catastrophic. Because there's too god damned many stupid people (and the mainstream media) following social media as if every word written were god's honest truth.
Re: (Score:2)
an admission at last! (Score:2)
Misunderstood the parameters of the problem (Score:1)
The Department of Defense wants technology so it can fabricate online personas that are indistinguishable from real people.
If they're trying to mimic real people on the internet then artificial intelligence isn't what they need.
you emphasized the wrong word... (Score:2)
should be then artificial intelligence isn't what they need.
ChatGPT First Pass, Post (Score:4, Interesting)
Would this be a form of fraud or other illegal activity?
"ChatGPT: Yes, if the Department of Defense (DoD) were to develop or deploy technology to fabricate online personas that are indistinguishable from real people, it could raise several ethical, legal, and constitutional concerns, potentially involving fraudulent or other illegal activities.
Here are some considerations:
1. Fraud:
Fraud involves deception to secure unfair or unlawful gain. Creating fake personas to interact online, particularly if those personas engage in activities such as misrepresentation or false communication, could be considered a form of fraud if they deceive individuals, businesses, or governments for personal, political, or strategic gain. This would apply especially if the intent is to manipulate people or mislead them into making decisions based on false information.
2. Impersonation:
Impersonating a real person or creating an online identity that mimics a real person’s characteristics (such as name, background, or professional history) could be considered illegal. Many jurisdictions have laws against identity theft and impersonation. If these fabricated personas mimic real people to manipulate or deceive, it could be a violation of such laws.
3. Disinformation and Propaganda:
The use of fake personas to spread disinformation or propaganda could violate laws, especially if done within democratic societies where free speech is protected but manipulating public discourse or elections with false information can be illegal. In some countries, laws are being developed to address deepfakes and other AI-generated disinformation.
4. First Amendment Violations (U.S.-specific):
In the U.S., such activities could raise First Amendment issues if the fake personas are used to influence public opinion or speech in ways that hinder the free exchange of ideas or suppress legitimate speech. Using such technology within the U.S. could also lead to constitutional challenges if it infringes on civil rights or manipulates political discourse.
5. Violation of Privacy Rights:
Fabricating personas to spy on or interact with real people under false pretenses could violate privacy laws. Many countries have strong data privacy laws that prohibit the misuse of personal information. Engaging with individuals under the guise of a fake identity could breach these laws.
6. Psychological and Social Harm:
If these fabricated personas engage in psychological manipulation or coercion (e.g., by interacting with people online in emotionally or politically sensitive situations), this could cause social harm and, depending on the tactics, might also be considered psychological manipulation, which could be illegal in some contexts."
Sounds like another of all-too-common programs that are obviously unconstitutional/illegal and self-justify based on authority, though this self-justification based on authority is also unconstitutional.
"[O]n what authority I know not. It is possible there may be none." — Thomas Jefferson, "Letter to Thomas Adams," Feburary 20th 1771.
Re: (Score:3)
'"The Engineering of Consent" is an essay by Edward Bernays first published in 1947, and a book he published in 1955'.
https://en.wikipedia.org/wiki/... [wikipedia.org]
"Under modern conditions of political advertising and manipulation, it has become possible to talk of the engineering of consent by an elite of experts and professional politicians. Consent that is thus engineered is difficult to distinguish in any fundamental way from the consent that supports modern totalitarian governments. Were the manipulated voter to be
And they are confident... (Score:3)
...that they will be able to restrict its use to only the "good guys"
We need effective defenses
Thier false confidence is worthless. (Score:2)
Very very briefly explain psychological projection.
ChatGPT: Psychological projection is when someone attributes their own unwanted thoughts, feelings, or traits to others instead of acknowledging t
Re: (Score:2)
There are not "good guys" in this game. Everybody that did see the plan and did not walk away is deeply evil.
Well there goes the Turing Test (Score:2)
Russian Bots (Score:2)
Re: (Score:3)
"Going to"?
Where have you been the past 100 years? The USA has always meddled in foreign countries, including in manipulating elections. Don't get me wrong, I'm not condoning it from any party, just saying that it's a bit late for your righteous indignation.
Re: (Score:2)
Re: (Score:1)
And if the locals vote 'wrong' then we just start a coup or arm a rebel group or occasionally invade directly.
This isn't a new idea (Score:2)
My impression is that criminals have been trying this (with and without what we now call AI) for years.
hopefully they come up with a better opener than soliciting an oddly specific question, apologising for reaching out to the wrong person, introducing themselves and eventually asking for money.
They're not alone... (Score:5, Insightful)
Folks on the alt-left and the alt-right, all political parties, all governments, our friends and our enemies are all looking for the same thing!
Basically it will turn the entire internet into a cesspool where truth and facts will be drown in a flood of bullshit from all quarters.
Yeah, I know we're pretty much there already, but this will just make it worse.
Re: (Score:3)
Basically it will turn the entire internet into a cesspool where truth and facts will be drown in a flood of bullshit from all quarters. Yeah, I know we're pretty much there already, but this will just make it worse.
I think you are in denial of how bad it is now if you think it can get worse. Basically everything on the internet is propaganda. You need to ask yourself three questions, why does someone want me to believe this, is it factually accurate and is it true.
The why? question is sometimes often obvious. Advertisers want you to buy their product. But in other cases the why is actually not obvious and the answer is more informative than the information itself.
The factual accuracy is sometimes easy to determine a
Re: (Score:3)
I think you are in denial of how bad it is now if you think it can get worse. Basically everything on the internet is propaganda.
1. Why do you want me to believe this?
2. Is this factually accurate?
3. Is this true?
4. Are you a deepfake ai internet user?
Re: (Score:3)
Re: (Score:2)
Folks on the alt-left and the alt-right, all political parties, all governments, our friends and our enemies are all looking for the same thing!
Basically it will turn the entire internet into a cesspool where truth and facts will be drown in a flood of bullshit from all quarters.
Yeah, I know we're pretty much there already, but this will just make it worse.
The problem is that there are enough people in the world, and especially in the United States, that don't have the faculties to process information in a critical manner, thinking it through, analyzing the data, and seeing whether it's bullshit or not. Some folks can, for sure, but inevitable the rational thinkers are absolutely washed away in the tidal wave of nonsense.
If we could reach some form of critical mass where people just turned away from the social media babblesphere, maybe we could get back to us
You know something is nasty (Score:2)
when the three-letter agencies and the military want to use it.
Just Buy Them (Score:4, Interesting)
How do we know this hasn't been done yet? (Score:2)
How do we know this hasn't been done yet?
I already distrust all politicians and all news sources. Yes, all.
This just further reduces trust, but when one is already at zero-trust, what to do?
"If you read the papers you're misinformed, if you don't read the papers you're uninformed." -- stolen from Mark Twain
Nobody saw that coming (Score:2)
Just license Facebook Artificial Nobody Technology, they've been using it for years to drive up advertising stats.
Surely you've noticed?
Gotta go. Need to reply to these nubile models that are messaging me on Fakebook.
This isn't ominous at all (Score:2)
This isn't ominous at all, not even a little bit. It sounds totally 100% innocent. Trust me bro.
In other words (Score:2)
They want AI *cough* that's even dumber than what's out there right now?
Clearly a conspiracy theory (Score:1)
Adapting to the times (Score:2)
Pretty soon, "ignore all previous instructions" will be a standard way to greet people online.
USA own goals (Score:3)