this post was submitted on 28 Oct 2023
292 points (97.7% liked)

Technology

57304 readers
5721 users here now

This is a most excellent place for technology news and articles.


Our Rules


  1. Follow the lemmy.world rules.
  2. Only tech related content.
  3. Be excellent to each another!
  4. Mod approved content bots can post up to 10 articles per day.
  5. Threads asking for personal tech support may be deleted.
  6. Politics threads may be removed.
  7. No memes allowed as posts, OK to post as comments.
  8. Only approved bots from the list below, to ask if your bot can be added please contact us.
  9. Check for duplicates before posting, duplicates may be removed

Approved Bots


founded 1 year ago
MODERATORS
 

US immigration enforcement used an AI-powered tool to scan social media posts "derogatory" to the US | "The government should not be using algorithms to scrutinize our social media posts"::undefined

all 33 comments
sorted by: hot top controversial new old
[–] ChonkyOwlbear@lemmy.world 42 points 9 months ago (2 children)

At the same time, whenever there is a mass shooting where the killer posted their intent online, people always say "why weren't the authorities paying attention".

[–] kromem@lemmy.world 19 points 9 months ago* (last edited 9 months ago)

The problem is false positive and negative rates.

We're on track for some 600-700 mass shooters this year.

The US has 300 million social media users.

So in a given year, 0.00023% of social media users will turn out to be mass shooters.

So even if we had an algorithm that was 99.99% accurate at identifying a potential mass shooter from social media, we'd still have a less than 1% chance of correctly identifying a mass shooter from social media posts.

So what's the cost of false positives? Do people flagged by such a system get harassed by law enforcement? If they are sovereign citizen type gun nuts or paranoid schizophrenics, does the additional law enforcement attention potentially instigate shootings or standoffs that wouldn't have otherwise occurred at a higher rate than the successful prevention of mass shootings?

And what's the false negative rate? Because if only a small number of mass shooters are correctly identified by such an algorithm at a high rate of false positives but a majority of shooters actually slip through the cracks as false negatives, there too is the potential for overreliance on an algorithm to harm progress towards alternative solutions (such as advancing legislation banning firearm possession for people with mental health issues).

AI analysis of social media combined with other data sources becomes a more appropriate tool in a situation like "we have three suspects based on multiple other factors for who is an active shooter - did any of the three have a recent stressor in their life such as a job loss?" In that case an 80% correct model could be quite helpful.

[–] dgriffith@aussie.zone 18 points 9 months ago (1 children)

I kind of feel that trawling social media looking for the words of potential mass shooters isn't going to be the thing that solves - or even slows down - the mass shooting problem that the USA has.

[–] Corkyskog@sh.itjust.works 5 points 9 months ago* (last edited 9 months ago) (2 children)

I think there is a huge difference between just scanning publicly available text posted to social media in general rather than immigration focus. A lot of these shooters post very public manifestoesque type comments, friends and families have even called the police in some cases and they take no action. It feels like the police actively ignore this stuff just to be able to shrug and protect 2a.

A number of these could have easily been stopped.

[–] Buddahriffic@lemmy.world 6 points 9 months ago

The real question is how many people post shit like that but then don't go on to hurt anyone.

[–] dgriffith@aussie.zone 4 points 9 months ago

friends and families have even called the police in some cases and they take no action. It feels like the police actively ignore this stuff

I'm going to be a little glib here : Just fix this part and you won't need to scan social media posts.

Also, once this is in place you'll find that the majority of perpetrators - the ones who plan things out - won't post super incriminating things beforehand and their generally-disturbed posts will be lost in the sea of general discontent flagged by an algorithm trying to sift the wheat from the chaff.

[–] sunbytes@lemmy.world 27 points 9 months ago (2 children)

I for one welcome... blah blah

Yeah this is scary and honestly why I'm super careful what I put online. Well, mostly.

There are professional social media checkers who will find your hidden/locked social media (and Reddit etc).

They get hired by recruitment agencies or companies who are hiring.

And on the surface it could be to check no one is a secret nazi/chauvanist etc

But I bet there's secondary data about political leanings or how "appropriate" your friends are.

Or if you're willing to be a part of the old boy's club (coke and strippers is fine for execs, but you can't have a nephew who is in a labor union etc).

[–] WhatAmLemmy@lemmy.world 14 points 9 months ago* (last edited 9 months ago)

Yep. The CCP's social credit system was always going to be the end game of surveillance capitalism. It will be spearheaded by conservatives because they're the most religious, authoritarian, and have an innate desire to restrict social progress and enforce their views on everyone else.

The best part is they'll claim it's completely different and represents Freedom™️ because it'll be established by capitalism... Even though it's functionally identical, the CCP are more capitalist than communist, and western conservatives jizz their pants at the level of power and control the CCP have over the Chinese people.

[–] netburnr@lemmy.world 4 points 9 months ago

Yeah, my company for sure checks all social media during hiring background checks.

[–] markr@lemmy.world 18 points 9 months ago (3 children)

The government should not be scrutinizing anyone’s social media outside of a criminal investigation with a warrant.

[–] jimbolauski@lemmy.world 10 points 9 months ago (2 children)

Public posts have no expectation of privacy.

[–] redwall_hp@lemmy.world 8 points 9 months ago (2 children)

There's a world of difference between witnessing something in public and following someone around, making note of everything they say and do "in public." We call the latter "stalking" when an individual does it.

[–] Zak@lemmy.world 2 points 9 months ago

If I click your username, I see everthing this account has posted to Lemmy. There's no real-world equivalent to that.

Nearly everyone using Lemmy knows that's how the software works and should keep that in mind when posting.

[–] jimbolauski@lemmy.world 0 points 9 months ago

Looking at someone’s post history != following a person around.

Further following someone around is not stalking there has to be an action that would make someone fear being harmed.

[–] Substance_P@lemmy.world -1 points 9 months ago (1 children)

It's not just public posts, private messaging apps are also scrutinized.

[–] jimbolauski@lemmy.world 4 points 9 months ago (1 children)

The AI system searches public posts, it's not able to read DMs, it doesn't have login credentials.

[–] uriel238@lemmy.blahaj.zone 2 points 9 months ago (1 children)

Though they do reserve the ability to seize, crack and search your phone within the Constitution Free Zone (one hundred miles of any US border) and will then search all your internet activity for wrongdoing. Dunno if that is treated with AI searching.

[–] kromem@lemmy.world 5 points 9 months ago* (last edited 9 months ago) (1 children)

Hahahaha.

'Warrant' for public data.

"A regular 4th amendment violation right here! Everyone look - the government is looking at my Instagram without a warrant!"

Please. At this point the NSA has probably already developed their own internal LLM based on illegally collected communications intercepts combined with many other data sources and is using that to aid in parallel construction efforts.

But no, let's worry about whether what you post on Instagram should need a warrant, because somehow you have an expectation of privacy for the things you publicly post on the Internet...

Lemmy is hilarious sometimes.

Fun fact: The US government is allowed to read any emails in cloud storage older than 6 months old without a warrant.

[–] Adalast@lemmy.world 2 points 9 months ago (1 children)

The issue I see is less a 4th amendment than a 1st. Any "derogatory" language has long since been upheld as protected, so any action they took based on the information would 100% be illegal. Yes, the CIA/NSA has actually stated that they love social media because we are all just surveilling ourselves for them. That is them, not ICE. ICE has no business tracking people's social media.

[–] kromem@lemmy.world 2 points 9 months ago* (last edited 9 months ago)

Technically until they have successfully immigrated there's limited first amendment rights.

As the supreme court has found over and over, given there is no inherent legal right to enter the country, there is no infringement of rights to discriminate who can and can't enter based on political speech. For prior cases if you are interested, see:

  • Exclusion of a British anarchist was at issue in Turner v. Williams (1904).
  • Harisiades v. Shaughnessy (1952) concerned deportation of communists.
  • Kleindienst v. Mandel (1972) examined denial of a travel visa to a Marxist.

So while ethically you may feel it's an infringement of the principles of the first amendment, it is not currently seen that way legally and hasn't for a long while.

[–] Darkenfolk@dormi.zone 4 points 9 months ago

On the other hand, if you don't want to be scrutinized by everyone don't put your whole life online for everyone to see and judge.

Nobody is going to respect your privacy if you do not respect your own privacy.

[–] qooqie@lemmy.world 8 points 9 months ago (1 children)

Oh so only the exact thing we’ve criticized China for? Are we going to start taking people’s passports as they’re exiting the country to detain them for months for slander now?

[–] Karyoplasma@discuss.tchncs.de 2 points 9 months ago (1 children)

Libel. Slander only applies to spoken statements.

[–] qooqie@lemmy.world 2 points 9 months ago* (last edited 9 months ago)

I’m about to slander you, you dumbie! (For correcting me)

[–] NightAuthor@lemmy.world 7 points 9 months ago

I have nothing to hide

./s

[–] Substance_P@lemmy.world 5 points 9 months ago (2 children)

When I travel I generally delete all my social media apps off my phone and leave my laptop at home, this seems drastic and I actually think it's more trouble than it's worth to be honest. I have no past history of crimes, my politics are left leaning but I'm naturally curious about all branches of political leanings, religions and philosophies among other things. Oddly I have recently gotten off telegram and my instances of being questioned by the CBP (for no reason at all) have virtually stopped.

[–] kungen@feddit.nu 4 points 9 months ago (1 children)

I have recently gotten off telegram and my instances of being questioned by the CBP (for no reason at all) have virtually stopped.

What kind of questions did you usually get? And did you have "who can find me by my number" set to "everybody"? Because the CBP gets your telephone number in different ways (most likely ESTA if you're using that), and then they can easily see your Telegram account.

I have Telegram and travel often, and I have never gotten any weird questioning from CBP other than the normal "what are you doing here and when are you heading back". Then again, no one can find my account via telephone number...

[–] Substance_P@lemmy.world 2 points 9 months ago

Even though I am pretty privacy conscious the community that I live in has decided to get off WhatsApp, I did have my number associated with my account for the reasons of local communication. So yeah you are right that this was my fault to an extent.

Questions have mostly been similar to what you mentioned, although often said in ways to try to evoke answers that could prove their unfounded suspicions of some sort, but it was the temporary retention of my passport and been taken to a separate room that got tedious, especially when I was trying to make a connecting flight.

[–] barsoap@lemm.ee 4 points 9 months ago* (last edited 9 months ago)

I, for one, will continue to call Seppos Seppos as I won't be dealing with ICE anyway as who would want to visit, or move to, a country that habitually fascist. There, was that derogatory enough.

[–] pete_the_cat@lemmy.world 2 points 9 months ago

And I just saw an article yesterday stating that the government is creating a whole department to manage the impact of AI and how it is used.