DHS Has Spent Millions On an AI Surveillance Tool That Scans For 'Sentiment and Emotion' (404media.co) 50
New submitter Slash_Account_Dot shares a report from 404 Media, a new independent media company founded by technology journalists Jason Koebler, Emanuel Maiberg, Samantha Cole, and Joseph Cox: Customs and Border Protection (CBP), part of the Department of Homeland Security, has bought millions of dollars worth of software from a company that uses artificial intelligence to detect "sentiment and emotion" in online posts, according to a cache of documents obtained by 404 Media. CBP told 404 Media it is using technology to analyze open source information related to inbound and outbound travelers who the agency believes may threaten public safety, national security, or lawful trade and travel. In this case, the specific company called Fivecast also offers "AI-enabled" object recognition in images and video, and detection of "risk terms and phrases" across multiple languages, according to one of the documents.
Marketing materials promote the software's ability to provide targeted data collection from big social platforms like Facebook and Reddit, but also specifically names smaller communities like 4chan, 8kun, and Gab. To demonstrate its functionality, Fivecast promotional materials explain how the software was able to track social media posts and related Persons-of-Interest starting with just "basic bio details" from a New York Times Magazine article about members of the far-right paramilitary Boogaloo movement. 404 Media also obtained leaked audio of a Fivecast employee explaining how the tool could be used against trafficking networks or propaganda operations. The news signals CBP's continued use of artificial intelligence in its monitoring of travelers and targets, which can include U.S. citizens. This latest news shows that CBP has deployed multiple AI-powered systems, and provides insight into what exactly these tools claim to be capable of while raising questions about their accuracy and utility. "CBP should not be secretly buying and deploying tools that rely on junk science to scrutinize people's social media posts, claim to analyze their emotions, and identify purported 'risks,'" said Patrick Toomey, deputy director of the ACLU's National Security Project. "The public knows far too little about CBP's Counter Network Division, but what we do know paints a disturbing picture of an agency with few rules and access to an ocean of sensitive personal data about Americans. The potential for abuse is immense."
Marketing materials promote the software's ability to provide targeted data collection from big social platforms like Facebook and Reddit, but also specifically names smaller communities like 4chan, 8kun, and Gab. To demonstrate its functionality, Fivecast promotional materials explain how the software was able to track social media posts and related Persons-of-Interest starting with just "basic bio details" from a New York Times Magazine article about members of the far-right paramilitary Boogaloo movement. 404 Media also obtained leaked audio of a Fivecast employee explaining how the tool could be used against trafficking networks or propaganda operations. The news signals CBP's continued use of artificial intelligence in its monitoring of travelers and targets, which can include U.S. citizens. This latest news shows that CBP has deployed multiple AI-powered systems, and provides insight into what exactly these tools claim to be capable of while raising questions about their accuracy and utility. "CBP should not be secretly buying and deploying tools that rely on junk science to scrutinize people's social media posts, claim to analyze their emotions, and identify purported 'risks,'" said Patrick Toomey, deputy director of the ACLU's National Security Project. "The public knows far too little about CBP's Counter Network Division, but what we do know paints a disturbing picture of an agency with few rules and access to an ocean of sensitive personal data about Americans. The potential for abuse is immense."
Doesn't matter if it works (Score:2, Interesting)
Only matters where they try to apply it.
At this point, you can figure that ~50% of voting Americans are disgruntled to some degree, and as of 2020 only ~60% of eligible voters bothered to show up - so you're statistically guaranteed to have ~30% of all Americans rank poorly according this sort of tool.
Being irritated isn't a crime. Hell, posting unrealistic threats isn't even a crime.
Makes me wonder why this tool even has any face value.
Not for Airports (Score:2)
Only matters where they try to apply it.
Well, I doubt it will be much use at an airport given air travel today. How will it distinguish between an angry, suicidal terrorist and the average passenger who is mad at all the crap they have to put up with and is slowly losing the will to live?
Re:Not for Airports (Score:5, Funny)
The terrorist will be the calm one, knowing that at least it's going to be over soon and that he will escape this travel hell.
Re:Doesn't matter if it works (Score:4, Interesting)
It matters because it doesn't work ..because it's unreliable and not admissible in court ...
20% of US citizens live near enough to a 'port' to be monitored, and a large proportion of the rest travel, and so can also be surveilled, but jurisdiction limits haven't bothered them for years ...
Re: (Score:2)
>It matters because it doesn't work ..because it's unreliable and not admissible in court ...
But it does work. Sentiment analysis is a real thing, and used by people who make money with it. Serious money. Get with the times, grandpa!
As far as being admissible in court, you would need to say that they illegally had access to the source material, not that they used a sentiment analyzer. Sentiment analysis would flag someone for further scrutiny. It's their damning text messages, social media posts, em
Re: (Score:2)
It matters because it doesn't work ..because it's unreliable and not admissible in court ...
It does not need to be admissible in court. It just needs to generate probable cause on demand, like a drug sniffing dog.
Re: (Score:1)
Re: (Score:2)
This is the problem with our government. Its so large it defies any real accountability.
Exactly nobody can tell you what is in the budget. Even the people like congressional aides whose very business it is to read the bills, not pass them to see what is in them, probably can't define half the terms for things.
Government it is a poor fit for so much of what its trying to do. While it can and must walk and chew bubble gum so to speak, it needs to cut the agenda down small enough number of head line items tha
Re: (Score:1)
It's fine when algorithms are used to "measure" and "guide" consumer decisions or to survail the common folk yet that enthusiasm fades when a decision comes from the top down to limit YOUR freedoms. The fact is, if the most privileged among us weren't so immoral and unprincipled all this AI control would still be in the background with the common people never knowing they are victims of a privileged conspiracy. Who knows all the seeded distrust would even have started a proper war by now thus fueling the d
Re: (Score:2)
>Hell, posting unrealistic threats isn't even a crime.
What's an unrealistic threat these days?
Saying you are going to harm or kill someone isn't unrealistic at all these days, even if you use hyperbole is saying so. There are *plenty* of deranged people would would like the ability to get away with doing just about anything someone has threatened.
I suppose something like "I'm going to swallow the Earth in a giant black hole!" would quality.
But run of the mill "unrealistic" threats because you believe th
Re: (Score:2)
It's almost like there is a ton of documentation and prior case history that helps us determine if a threat qualifies as a crime. :)
https://www.criminaldefenselaw... [criminalde...lawyer.com]
Re: (Score:2)
Eh, why would they use this to train the AI? That makes too much sense!
Sentiment And Emotion Hunting (Score:2)
They want to train on Sentiment And Emotion Hunting?
I assume this means they've set their AI to train on trolling?
--
Flattery and insults raise the same question: What do you want? - Mason Cooley
Re: (Score:1)
I'm sure sentiment is collectively much more subtle. I guess the choice now is DHS? otherwise a facebook like entity? or aTicTok?
Adventures in skimming TFHeadline (Score:2)
Read this as "DHHS Has Spent Millions On an AI Surveillance Tool That Scans For 'Sentience and Emotion'" and was wondering why the Department of Health and Human Services was so concerned about detecting automated messages.
Bait Much? (Score:1)
I feel as though a trolling has been befallen me.
Nothing new (Score:2)
Sentiment analysis had buzz nearly two decades ago [google.com]. Does anyone believe the internet panopticons haven't been doing sentiment analysis at least since then?
Re: (Score:2)
Sentiment analysis had buzz nearly two decades ago [google.com]. Does anyone believe the internet panopticons haven't been doing sentiment analysis at least since then?
I'm sure I read about it right here on Slashdot two decades ago. Remember, that was peak hysteria about "terror" and peak jingoism about starting wars because of "WMDs." The USA had a shiny new TSA who were desperate to justify their funding by chucking dollars at whatever shiny pseudoscience caught their eye.
What I would really like to see is a study that quantifies how many dollars of GDP and person-years of life have been wasted in TSA queues, missed flights etc since September 11, 2001.
Voight-Kampff (Score:4, Interesting)
The DHS do know that was fiction, right?
The trouble with measuring emotions is that there are no broadly agreed external correlations with inner feelings (different people show their feelings in different ways) & that it's usually very, very difficult to determine the cause of those feelings, e.g. Is she feeling awkward about the question the agent just asked or is that he's just creepy?
The NSA & DHS (Score:2)
If I recall, the NSA & DHS started doing this 10 years ago: Has it worked? Well, they didn't search for "a gun and kill people" and the USA has suffered a number of mass murders over the years.
Translation: CPB has already decided who they need to arrest.
Scan this emotion: (Score:1, Insightful)
Re: (Score:2)
Hey DHS, DEA, FBI, CIA... we all hate your fucking guts. You corrupt and evil pieces of shit. You will be purged from human civilization. We will not tolerate your kind any more.
Sentiment and emotion scan complete. Risk terms and phrases detected. Initiating resolution process omega three niner.
Re: Scan this emotion: (Score:2)
Your order of 9x Carls Jr fish sandwiches has been submitted
What is the False Positive Rate? (Score:2)
Re: (Score:2)
Money has been spent. Where is the post implementation report stating the false positives and false negative numbers?
They do not track false positives from drug sniffing dogs, so why would they track them for this?
Re: (Score:2)
Unmonitored conversations verboten (Score:2)
With all this monitoring, one day there will be a terror event where the perpetrator did not use social media at al. That will be the day politicians will ban unmonitored conversations.
Re: (Score:3)
No, that will be the day those without social media accounts won't be allowed to fly.
impersonation (Score:2)
Automation example (Score:2)
Trump, P01135809 mug shot emotion (Score:2)
https://en.wikipeda.org/wiki/M... [wikipeda.org]
Meanwhile (Score:3)
DHS pisses away millions on a boondoggle project (Score:2)
Thought-Crime (Score:2)
And emotional crime. This is what they want to establish.
Come to think of, Trump looks a bit demonic in his mug-shot. What would this software make of that? Potential terrorist?
Fantastic! (Score:2)
Detect hate?
Give them gun-ads.
Detect suicidal feelings?
Give-them gun-ads.
Detect frustration? ...
Give-them gun-ads.
Troubling because the DHS-CBP are morons (Score:2, Insightful)
Combined with Biden's ill fated Ministry of Truth (Score:1)
Re: (Score:2)