an actually useful feature for Apple Intelligence would be to identify these incredibly obvious scams from the half-dozen obvious tells it's a scam and then alert the user in the summary "hey this is a scam" so obviously it doesn't do that
Comments
Log in with your Bluesky account to leave a comment
It is wild how real that website looks when you click on it. Stealing the logos and everything. This is probably costing people millions of dollars and it is a problem worth solving.
Got a new Pixel recently and it does it pretty well with both calls and texts. I can spam calls on the spam list for pixel. It also flagged a text back from a car dealer that I reached out to as spam, which is a win.
"AI bad" amirite liberal. what, you want resources going towards AI to be used towards achieving useful applications? don't you know there was a famine in russia in 1935?
I hope that people who have a large number of followers will help the Gazans’ campaigns to achieve their goals. This is the last hope for our lives, my friends.🙏🙏🙏
ive said for years the reason they (feds, telecomms) don't completely wipe out these scams is because It's Too Much Effort and it would require a revamp of how anti-fraud laws work, thus indicting official channels
I reply to them with variations of "They're the POST OFFICE, dumbass, they have my ADDRESS, not my cell phone number. This isn't how they do business" and then usually closing by suggesting they eat a bag of dicks or something similar.
I'm an online support specialist providing assistance with:
- Locating kidnappers
- Recovering funds lost to scams
- Monitoring social media activity for potential infidelity
- Mitigating blackmail threats
- Repairing and restoring damaged credit scores.
And my elderly mother, who is never expecting package delivery, falls for this scam text every time — sharing her name, address, phone number, and the same for 3 of her neighbors “just to make sure the package arrives.” 🤦♀️
What confuses me about these is that they're delivered via iMessage, which usually means there's an iPhone/iPad/Mac on the other end, but scam text blasts almost always come from spoofed numbers over SMS.
In Belgium we get official texts from numbers that are 4 digits, max. So scams are very obviously from cell phone numbers, which start with 04 and are 10 digits long. Does the US not have phone numbers that are reserved for legitimate, more official notices?
There are typically 5-digit numbers that are used for notifications and responses, but it's not reserved for the government, just more expensive.
And cell phone numbers don't have a dedicated area code;
they share those with landlines.
A non-US country code is often a hint, though.
anything referring to whatsapp in the US is basically guaranteed to be a scam because no one uses it here unless they are communicating with a person in a country where whatsapp is popular
I think my cousins use WhatsApp; whatever tool they're using does group chats or group audio calls or something,
and isn't Signal. Might be Telegram or one of the others instead.
my shitty old samsung AG15 tags these texts as spam and as far as i know there is no AI functionality built into its version of android OS, so oof apple
I know that it’s not a real text from USPS because:
A) they didn’t use a short code, which most businesses use, and would probably be used by the USPS for texts. The scams use a full 10 digit number or an email.
B) they thanked me. USPS is never friendly, nor are they mean. They are neutral.
It doesn’t feel like AI will solve spam. In general any unwanted communication will tend to be spam. Some kind of mutual agreement to exchange messages will ultimately be the solution
Just got one of these this morning. They make group chats, post the scam, then kick you, but you can still see it. I think they do the group chat thing so the report spam option gets confused and reports one of the other numbers and not the one who actually made the group/scam?
Honestly despite my feeling on AI overuse this is one field that I think that LLM's would be extremely good at. Spam filtering, profanity filtering, etc. Instead of regex type pattern matching it gets processed by a LLM and you end up with a much more robust filter (in theory)
It would be nice if our super-wealthy, advanced, high-tech, innovative society could come up with some way to stop *every* method of communication from filling up with so much garbage that it becomes utterly useless.
I blame Diane, whoever she was. She lived in very Republican Ocean County, NJ; ... had my number before I did but I've had my cell number since 2003, so I figure she probably died.
Folks still call/text asking about her property or donations to the local GOP committee.
I haven't paid enough attention to those to notice if they're
always from the same +63 phone number in the Philippines,
and I don't get it often enough to bother blocking,
but yeah, Apple Intelligence could probably do that on the phone itself without needing Apple servers to monitor every text.
The system as it stands is pretty much entirely powered by folks hitting “report junk”, which I’m not gonna claim is the best implementation but does help explain why it doesn’t flag things that feel like they should obviously be flagged.
That being said, holy shit you should hear people’s reactions when their messages ARE being flagged (or when they catch a flag/10 day SMS sending block for using their phone for “non-consumer messaging”
Realtors screaming that they need text messages for their job while fully admitting that they sent a bajillion duplicate text messages is kinda hilarious in its own right.
Shoutout to Pixel phones that already blocks those kinds of spam. And you can block by country code too if you feel that necessary. Apple is behind a bit on the spam blocking/phone assistant gatekeepers that I appreciate so much on Google phones.
I actually got a call from USPS asking me if I was expecting a package from Mexico. I said “yes”. The guy said “what’s in the package”? I said “cocaine”. He told me to shove the cocaine up my ass and then hung up :(
I used to frequently answer obvious spam calls as "AT&T Security",
which was technically correct even though what I did there was computer security development, not fraud detection,
but they were calling on my work cellphone.
like, the algorithms they have running could easily, easily identify any of this. Same with like, dating app bots, Instagram scammers, etc. They simply don't want to
i have a hunch there’s a lot of money to be made in allowing spammers to do what they do, for social media and cell providers alike. otherwise something would have been done by now.
oh yes, organizations like the various "Police Benevolent Society" etc. it's how they make their money, and there's absolutely a feedback loop between the FOP and politicians
also seems pretty easy to see a Filipino number texted "USPS" with a suspicious link and at least throw a ⚠️, particularly if you've never texted anyone from the Philippines, but what do I know
A decade or so ago I got a call from "The IRS", but from a Rochester NY telephone number (turned out to be a Magic Jack VOIP gateway); since I'm in California, if the real IRS called it'd be from Fresno or DC.
(I do miss having all the cool infrastructure lookup features from working at a telco.)
Right? The first time whit one of these I had just ordered some throw pillow covers from an Etsy shop based in Turkey. So I wondered if it was legit for a beat.
this is probably why they don't do it, and if it's not, it's why I'd ask them not to do it hastily - labeling "this might be a scam" on only some scams seems worse than doing nothing, unless you think about how to get people to realize no warning does not mean no scam
*that said* this is a classic use of even very shallow statistical language matching algos, modern deep ones would be great if they weren't dramatically more gullible than humans (which is saying a lot)
This feature does exist and I've been using it for a couple of years now.
I turned on "Silence Junk Callers" turned on. In my recents tab on my phone app, I have a "scam likely" notice on scam calls and a "scam likely" notice on scam texts like the one you got.
I actually fell for one recently because it was Christmas and I had 8 billion packages on the way and was jet lagged to hell. Google is usually pretty good about keeping scams out of my inbox but it slipped past. It did look extremely real.
That's basically because Caller ID was designed back when the only things that would EVER send it were other telcos that could obviously be trusted (even if some really couldn't),
and there aren't firewalls in the system to deal with that.
"Useful" and "Apple" are words that don't go well together. Most of the useful features on my iPhone are there because of independent tweak creators (I use a rootfull jailbreak). It's crazy how individuals do a better job of implementing useful features than Apple does.
I work in live events. I'm constantly getting calls and texts from unknown numbers for gigs. The phone is great at sorting the real ones from the spam. I would miss way too many if I did that.
Holy hell, that's crazy. I use and like some Apple products but when I read stuff like this they just come across as... primitive? Like just recently being able to arrange icons on your phone? Nuts.
yeah not having an app drawer i can sort alphabetically is irritating as well. overall i don’t mind iOS and it has some advantages, but some of the simplest things are just not an option.
Comments
- Locating kidnappers
- Recovering funds lost to scams
- Monitoring social media activity for potential infidelity
- Mitigating blackmail threats
- Repairing and restoring damaged credit scores.
And cell phone numbers don't have a dedicated area code;
they share those with landlines.
A non-US country code is often a hint, though.
and isn't Signal. Might be Telegram or one of the others instead.
A) they didn’t use a short code, which most businesses use, and would probably be used by the USPS for texts. The scams use a full 10 digit number or an email.
B) they thanked me. USPS is never friendly, nor are they mean. They are neutral.
Helping people loses money
Like, why is anything with a domain name with https://usps.com as a prefix not automatically labeled as “this is not the USPS”?
Only available in Canada, Denmark, France, the Netherlands, Portugal, Spain, Sweden, Switzerland, the United Kingdom and the contiguous United States
Folks still call/text asking about her property or donations to the local GOP committee.
always from the same +63 phone number in the Philippines,
and I don't get it often enough to bother blocking,
but yeah, Apple Intelligence could probably do that on the phone itself without needing Apple servers to monitor every text.
Less easy to do on device but genuinely would do an amazing amount of good to flag this stuff for people who otherwise wouldn’t know.
Glad to know I’m not the only one getting these at least.
which was technically correct even though what I did there was computer security development, not fraud detection,
but they were calling on my work cellphone.
Engagement is the 1st rail of social media--so many firms won't combat spam unless execs perceive it to be a critical issue.
Telecoms love data consumption, but not sure this is true as much as it used to be.
(I do miss having all the cool infrastructure lookup features from working at a telco.)
100%
Even better would be "hey, this is a scam so I'm just going to block the number for everyone"
I turned on "Silence Junk Callers" turned on. In my recents tab on my phone app, I have a "scam likely" notice on scam calls and a "scam likely" notice on scam texts like the one you got.
and there aren't firewalls in the system to deal with that.
https://support.apple.com/en-gb/guide/iphone/iph87abad19a/ios