Anti-fraud tools can't keep pace with robocall scammers
broadbandbreakfast.com
broadbandbreakfast.com
Result: All routes to non-cooperating networks get dropped within days to weeks and scam-calling stops being a lucrative business basically instantly.
You can't make a system that can force bad actors off the system that can't also be used to force politically undesirable actors off the system.
California has introduced bonding to telemarketing firms specifically. I feel that should apply at the carrier level, where networks carry a guaranteed bond, pay regular premiums on it, and are dinged for unwanted calls, with the proceeds being split among the called party and any third-party network(s) traversed by the calls. Downstream networks could seek compensation from ANY upstream network carrying the traffic regardless of whether or not they originated it.
This would both create a penalty for providing, or transiting, unsolicited calls, AND create an incentive for carriers / network providers themselves to pursue unsolicited traffic from their peers.
<https://oag.ca.gov/consumers/general/telreg>
________________________________
Notes:
1. See for example <https://toot.cat/@dredmorbius/111099306069523624>
And no specific instances or mechanisms detailed, to boot.
Thanks.
If your predecessor committed fraud, you were 100% personally responsible for it. You would then gather the evidence and sue your predecessor for your losses.
Scenario:
- Spamford places an unsolicited call to subscriber Alice initiating from MalTelCo, transiting carrier hops BunnTel1 and BunnTel2, to Alice's telco carrier, EndTelCo.
- Carriers MalTelCo, BunnTel1, BunnTel2,[1] and EndTelCo have all placed surety bonds, held by BondCo, to practice telephony operations within the jurisdiction (regional/national). The carriers are the Principals, BondCo is the Surety, and receiving subscribers (or telcos, see below) are the Obligees.[2]
- Unbonded carriers may have their traffic refused by peers. Peering to an unbonded carrier places the bond obligation on the receiving carrier.
- Alice flags the call as spam. A per-call surety of $100 is paid to Alice, and charged to EndTelCo against its BondCo contract. As an additional option the call may be flagged as fraud through the phone system, in which case it is automatically referred to LEO by EndTelCo. Obligation of surety is independent of any fraud finding and is based SOLELY on the unsolicited nature of the call.
- EndTelCo has the option of 1) eating the charge or 2) filing a claim against its peer, BunnTel2, the 2nd hop in the chain, which EndTelCo does.
- BunnTel1 similarly files a claim on BunnTel2.
- BunnTel2 files a claim on MalTelCo.
- MalTelCo now eats the claim (it's paid out by BondCo). MalTelCo may seek further compensation from Spamford, but that's Out Of Scope of the bonding / surety schema, and would be covered by MalTelCo's own terms of use.
- BondCo assesses risks and adjusts its surety rates correspondingly based on observed behaviours (and financial risks) of EndTelCo, BunnTel1, BunnTel2, and MalTelCo. If risks are excessive and no surety can be issued, MalTelCo is unbonded, and hence, decertified. Peers may now refuse traffic without penalty.
Note that no one carrier needs to know anything more about a call's routing than its own network boundary. If EndTelCo has no idea that BunnTel2 and MalTelCo were involved, it doesn't matter, because BunnTel1 is on the hook for passing on the call. Spoofing or falsifying records doesn't save you.
There are some questions over how this might be implemented, though generally:
- If Spamford and Alice are both subscribers to EndTelCo, then EndTelCo eats the surety, which is paid to Alice. There's no upstream. Moral: Telcos, don't spam your own customers.
- One thought is that the surety is split among telcos and the subscriber. Rather than just facing a potential cost, transiting and reciving-end-point carriers could see revenue by tracking and prosecuting unsolicited calls. This could include calls received by monitoring numbers set up strictly to assess unsolicited call activity directed to the network. This would mean that calls transiting multiple carriers would be subject to compounded surety claims ... and ... I think I'm OK with that.
- There would all but certainly be classes of calls which would be exempted from claims. Those should be very limited, preferably to government and specifically qualified emergency services only. No political exemptions, no non-profit / NGO exemptions.
- How often claims are settled and risks re-assessed is open for discussion. Daily might be too often, weekly or monthly seems most likely. Longer than that gives too much free-run for malevolent actors to operate.
________________________________
Notes:
1. "BunnTel", because bunnies hop.
2. For an overview of surety bonds, see <https://www.suretybondsdirect.com/educate/what-is-surety-bon...>.
The scammers who call me are perfectly obvious.
First, they tell me the company they are from (almost certainly a fake one -- could be easy to verify). Then they try to convince me that two years ago I have created an account on their website, they gave me some free money that was managed by an AI, and now I have a ton of money, and they need to send it to me (a completely bullshit story). Then they tell me that in order to get that money, I need to install a software, that I know happens to be a remote control software (no legitimate financial institution would ever do that).
There is no way to make this plausibly deniable. Especially the part about the need to install the remote control software... which is the entire point of the operation.
The delivery penalty applies to any unsolicited email, as determined by the recipient.
If also tagged as scam, those are further forwarded to law enforcement (state, national) for investigation.
Many US states are one-party regarding recording. Even in two-party states (CA, OR, WA, MT, IL, PA, MA, CT, NH, MD, DE, FL), disclosed recording and continuing a call will generally be construed as consent. If that's not the case, proposed state or national legislation could carve out exceptions as needed, and there'll likely need to be some legislation required anyway, so that's part of the process.
But shifting the fee element from fraud (one class of unsolicited call/text abuse) to undesired contact makes sorting when the fee applies far more evident, and eliminates a class of other objections (e.g., due process) from consideration.
If I were to steelman the concern, I'd look at a few related scenarios, say, where a subscriber is running a poorly-secured VOIP system and spammers hijack that to make calls. I'll ... get to that.
First: the scenario here is phone systems, not email, so the traffic would be voice calls, possibly texts. That said, I'll consider your question as if it was calls and not newsletters.
I've given a more detailed breakdown of how I see a bonding system working here, you might want to read it before continuing with this comment: <https://news.ycombinator.com/item?id=49129679>.
Second: It's not subscribers who are on the hook for spam calls, but carriers. So Google isn't paying you, your carrier is paying you (via a Surety agent), with the option of recouping that penalty from an upstream carrier, if any. If you and Google are on the same carrier, and the call didn't transit any other networks, it's just you and your own telephony service provider (carrier).
A carrier might have its own TOU/TOS with its subscribers, and subscribers originating calls could and likely would attempt to recover abuse costs if they were incurred. That subscriber (say, Google) might also have its own TOU/TOS addressing the case of mis-reporting of authorised contacts. Those actions would be outside the bonding system itself. A party repeatedly abusing the system could be liable for other actions, including fraud or malicious damages.
Note that one of the interesting elements of bonding is that call origination becomes a risky activity for telcos. Presently, telcos are eager to enter such business, put few restrictions or obligations on their customers, and to prefer outbound traffic to inbound traffic. Under a bonding programme, this changes dramatically. Large-volume outbound traffic is a liability, where it does occur, it needs to be closely monitored and managed. Our poorly-secured VOIP system mentioned earlier would probably be subject to configuration/operation validation, pen testing, close monitoring for activity, and alerts/throttling if unexpected usage patterns emerge. All of this is now in the carrier's interest.
Third: The bonding scheme would be periodically settled among carriers. I've hand-waved how often this would occur, though somewhere between daily and monthly, with a shorter term more likely (malicious actors often shoot-and-scoot, we want to avoid that). So low-level skirmish actions such as you describe would tend to result in a net wash between carriers: claims on one would be balanced by claims on others.
Fourth: Just how Google came to communicate, what it's communicating, and the degree to which it's coercing, say, receipt of sales/marketing messages vs. strictly advisory messages tied to a service ... would probably have to be considered in a larger context, but would still be outside the bonding system itself.
Fifth: There's a model for how surety bonds and claims work in the State of California's syste. For a breakdown of that see: <https://www.jwsuretybonds.com/states/california/telemarketin...>.
A few other points:
- New relationships might be permitted through a contact request. This itself could be mediated by a known third party. Private individuals for personal contacts, commercial or governmental trusted parties in other cases. Effectively it's the social-introduction problem from before the age of mass communications brought forward. Such systems will have some friction (necessary to defeat spammers), but not so much friction that the system as a whole doesn't work.
- Bonding does not require strong KYC for small accounts. That is, the person wanting to buy a mobile phone and service anonymously could, but their device and service would be monitored for abuse. I expect a tiered system to emerge, with individuals, small, mid-sized, and large accounts, with increased controls and obligations proceeding with scale and/or capability.
- Generally, it's not individual accounts which are responsible for large volumes of outbound calls, absent an issue such as a proxy hijacking. Large outbound volumes will tend to be associated with known call- or data-centres, and can be managed as such.
- The goal is preservation of a general-availability, universally-accessible phone system. That works only if it is not systematically abused, which is presently the case. If trust in public-switched telephone networks, permitting direct-dial access to any other number, anywhere in the world, is lost, what we'll see is desertion to other options which serve specific individuals' and organisations' interest. We are already beginning to see this, though no one clear winner has emerged. Unfortunately, most of the alternatives are proprietary, though some federated networks might prove to be viable alternatives.
If you're acting fraudulently and at scale, there will likely be consequences.
If this happens occasionally, it's a feature of the system, and your reports effectively become a super-opt-out.
And if perhaps the problem does become sufficiently widespread, I'd be interested in seeing how you'd address it given one constraint: operating within the bonding/surety system I've suggested. And that the State of California and others have already enacted in some form.
In Canada at least, only one party has to consent to the recording
If you record your own phone calls that's not remotely illegal. Nor is it in my opinion unethical
This is why a company must inform you of the recording, but you do not have to inform them.
Still, as an individual wanting to record scam callers, you're in the clear to record calls that you are a part of
It's a common feature in better dashcams--you get say the 30 seconds before the thump that triggered it. Likewise, many high speed cameras that record some short action. They're actually always running, but dumping the end of the data, they only "record" when the trigger happens.
The bonding agent (the Surety) sets the bond rate based on the perceived risk of the venture.
Unbonded ventures are not permitted to operate. In a telco context, unbonded carriers would not be peered to other carriers.
Overview of how surety bonds work: <https://www.suretybondsdirect.com/educate/what-is-surety-bon...>.
California's present regulation: <https://oag.ca.gov/consumers/general/telreg>
After that that scam would result in the scammer paying the (e.g.) $10 upfront, the (up to; e.g.) 10% for every hop between networks and likely and transaction, processing and legal fees for the civil case (if he's not cooperating).
And he'll likely end up with criminal charges on top of that.
Also it's not like the scammer gets sued by some powerless private citizen authorities are likely to ignore. His opponent will be a telecom provider in his own jurisdiction.
"Cold calling" can either come from governmental organizations (IE, call from the police), or someone or some organization in your network can grant access to call, text, email, ect on their behalf.
The result is that SPAM has a chain of traceability, so you can then block people and organizations in your network who make unwanted calls.
Furthermore, regulation is needed so that your cable company can't pull dumb nonsense like saying "telemarking is required to use our service": This is where the SPAM button that you propose really comes in handy, because it holds accountable the fools who think that selling their customers' contact information is a good idea.
Now, my only worry is that I might get a call from someone who I need to talk to, but is blocked and I won't even know it. For example, what if I get an emergency call from police/fire/hospital and I won't even know it. I also tried using "Screen unknown caller" feature, but then, no one likes them and sometimes they think it is AI bot and they usually just hang up (which is great for spam caller but not for legitimate caller)
My only solution that I could think of is to have a dedicated phone line just for friends/family/work, and a second line for banking/shopping/utilities/everything else.
Scammers or spammers will never send an SMS with clarification that they wanted to call you.
No police/fire/hospital emergency outcome is going to hinge upon someone else answering a phone call or text messsage.
As a parent, the "block all numbers!" approach has always seemed incredibly naive to me.
Other organisations, institutions, and businesses too. HN discussion tends to focus on the consumer side of this, it's what most commenters have most familiarity with themselves, but you'd better believe that pretty much the entire phone customer base is fed up to there on this. Which puts the entire network at risk of defection, a risk that telcos have been talking publicly about for over a decade now:
[S]ince mid-2015, a consortium of engineers from phone carriers and others in the telecom industry have worked on a way to [stop call-spoofing], worried that spam phone calls could eventually endanger the whole system. “We’re getting to the point where nobody trusts the phone network,” says Jim McEachern, principal technologist at the Alliance for Telecommunications Industry Solutions (ATIS.) “When they stop trusting the phone network, they stop using it.”
<https://nymag.com/intelligencer/2018/05/how-to-stop-spam-rob...>
I've mentioned this on HN a few times: <https://news.ycombinator.com/item?id=21494300> <https://news.ycombinator.com/item?id=21542926> <https://news.ycombinator.com/item?id=28756827> <https://news.ycombinator.com/item?id=29003329> <https://news.ycombinator.com/item?id=31939562>.
Broadband Breakfast just addressed the issue as well in this Fediverse toot, calling out not just schools (subject of the legislation) but other affected entities: <https://mastodon.social/@BroadbandBreakfast/1169990755811584...>.
Your follow-up, here (<https://news.ycombinator.com/item?id=48938169>), was particularly insightful, and has influenced my thinking. Essentially: authentication / validation should happen out of band with phone number itself, for the reasons you've given.
I do suspect that for routing authentication, header-level signifiers should be reasonably useful, but for strong identity or authority attestation, they're not. That's ... a deeper problem, but also one which can be solved independently.
Oh, and I'd love to see that Dallas Morning News AT&T CEO interview story, if you could find it.
Would this be it?
"Watchdog Memo to AT&T's CEO: Didn't mean to get you in trouble", by Dave Lieber (July 8, 2016) <https://www.dallasnews.com/news/watchdog/2016/07/08/watchdog...>
AT&T took some bad press around that time, and in July 2016, the FCC, facing political pressure, and now sensing that the PR gods were now on their side, leaned on the major carriers to set up a 'task force' to get things moving. I remember sitting in the little room at the FCC while various telco industry grandees bloviated. We all agreed that ATIS and IETF should define technical standards since they'd already made a start (Jim M did ATIS, for example) and so everyone agreed that STIR/SHAKEN was going to be The Path.
The original STIR/SHAKEN assumptions were that the attestation would be carried throughout the network in a SIP header and that every SIP processing element would deal with it. Reality soon intruded in the form of ancient telco equipment that didn't have the headroom to process another bunch of bytes in every call setup message. Another idea was to have signing and verification done by SIP application servers, but that didn't pan out for the same reasons. (Forgive me, I'm going to toot my own horn here for a second:) I invented a scheme with my colleagues where the signing or verification was an HTTPS operation triggered by the SIP device at the network perimeter (a 'session border controller' in telco-speak) which made the problem more tractable. AT&T took my PowerPoint and submitted a stunningly close copy as their own to the FCC. I still have the slides somewhere. I was salty at the time but ah well, what can you do: my employer still made money off the product with other customers.
I've heard the "but we can't filter traffic at the network level" line from AT&T much more recently than that date. If there's any corporate learning that's occurred, it's not filtered through the ranks.
It seems as if sorting out the STIR/SHAKEN situation with smaller telcos is going to have to happen. How that happens is something I don't have much clarity on, though I suspect some degree of national regulation and assistance will be required in the US. Your work could be an element of that?
I've been ... unhappy ... with telecoms options for well over a decade now, and predicting the #DeathOfTelephony (one of my Fediverse topics/tags) for much of that time. It's coming about more slowly than I'd expected/hoped, but I'm seeing lots of strain. Enterprise/organisational frustration has been increasingly apparent over the past five years or so. Millennials and onward (as well as many Gen Xers) actively avoid voice calls. It's simply getting hard to connect to people. Then there's the privacy / surveillance / propaganda / manipulation elements.
I think what I'd like is a SIP trunk to the house, and manage remote calls over WiFi (possibly with a SIP phone, I've got an old GSM 2G I'm thinking of doing some hardware-hacking on, or I could buy something ready-made). Carry a standard mobile for emergencies, but avoid it if at all possible, and route most comms through the VOIP system with a bunch of bespoke rules.
If I were running a business there'd be other considerations. I'm ... glad I'm not, for the moment.
The extent to which something vaguely resembling that might be more widely used ... I'm not sure. I'm thinking through elements I'd like to incorporate, might post that later (probably to the Fediverse or elsewhere for now).
This fails the moment one of your "clean line" contacts downloads a sketchy app that sells all their contacts, and sells an updated list as your appear in their recent calls list - meaning it's not a one-time thing, it's for a few weeks after every time you call them.
Some carriers offer blocking all unknown calls from specified area codes. On Verizon that is "Neighborhood Filter", part of their "Call Filter" toolkit:
<https://www.verizon.com/support/knowledge-base-238154/>
<https://www.pcmag.com/news/verizons-neighborhood-filter-bloc...>
The way it's meant to be used is that the subscriber would block their own area code, and perhaps co-located overlays or neighbouring area codes. The limit for now is ten area codes.
Known or explicitly approved numbers are passed through.
The additional hack is that it's possible to request a number in any arbitrary area code, and spam and robocall rates vary tremendously across area codes. They're generally worst across the Deep South / Southeast (TX, OK, AR, AL, MS, TN, GA, SC, NC), and lowest in Alaska, Utah, Massachussetts, Washington, and North Dakota.
See:
"The Robocall Epidemic: Which states are hit hardest by spam calls?" (10 March 2026) <https://www.whistleout.com/CellPhones/Guides/robocall-epidem...>
"The Robocall Geography Tax: Why Your ZIP Code Determines Your Spam Reality" (23 October 2025) <https://www.karmacall.com/blog/southeast-spam-call-geography...>
I'd submitted the latter a few days ago: <https://news.ycombinator.com/item?id=49082473>.
Pick a low-spam, low-population state, request a number from one of its area codes (if not its one area code ;-), and then block all but known numbers from that area code.
When its someone legit I find that they actually go through the effort. I've had USPS explain who they were to my phone and I was able to answer them as a result.
It is insane that Android still does not have this option. It has to be some kind of software patent horseshit preventing them from adding the feature.
And further to that, the elderly are also the ones who cannot block unknown numbers, because doctors’ offices seem to have random numbers they call you from (they may have a pool of numbers but it’s not reasonable to add all of them to contacts). Blocking all unknowns would block these important health related calls as well.
________________________________
Notes:
1. Incidentally, the first and second largest-circulation magazines in the US now: <https://www.magazineline.com/blog/most-popular-magazines-in-...>.
This is a huge issue with scam/security awareness education. Too many legitimate orgs use the exact behaviors we tell people to avoid. Same thing with email, can't tell someone to never click links in emails when services keep relying on magic links, third-party notification domains, etc. SPF, DKIM, and DMARC do nothing because scammers will just typosquat.
In the phone number example, most of those numbers too are unlisted outbound numbers, you couldn't even google them to verify.
Half the battle is getting legitimate organizations to stop acting like scammers in the first place so that shady behavior becomes an obvious red flag again.
Which telcos?
I know this because I had to spend 6 months taking care of my dad while he was dying. We had so many medical people calling us and making appointments, and doing tests, etc etc. And on top of it about 10-15 spam calls a day and it's impossible to differentiate via caller ID. This was all while I was also working full time from home. It's a real hellscape, and these phone scammers are really having a negative effect on people's lives. It made my life miserable at one of the hardest times of my life, and I have nothing but contempt for these fuckers, and no empathy left for the people making the phone calls either. They're all scumbags and deserve time in prison.
see [1] for more on why stir/shaken hasnt helped as much as it ought to.
The US has the STIR/SHAKEN authentication protocols, but the telcos seem to have no financial are regulatory pressure to clamp down on all the sketchy intermediate carriers the scammers use to bypass them.
Rather then actually solve the problem, the US prefers to turn it into a business opportunity with the telcos selling subscriptions to blocking software and an entire ecosystem of moderately effective apps.
If 99.9% of us have internet phones, why aren't we using PKI, decentralized protocols, crowd sourced reputation to communicate instead of POTS phone numbers, SMS/iChat and relying on the carriers to police spam.
In much of the US, the push is on to retire the last twisted pair within a very few years, if not months. Utilisation rates are well into the single-digits and falling, which makes continued support quite expensive.
The problem for many holdouts is that alternatives fail to deliver reliability, or the equivalent of a site-centred service (as opposed to personal mobile devices). Residential VOIP is confusing from the subscriber's perspective, and the telcos aren't making choosing options much easier.
SMS itself is highly problematic, as it's grossly insecure, unreliable, and very subject to surveillance and other abuse. Secure chat alternatives tend to be proprietary (e.g., RCS, effectively specific to Apple and Google, see: <https://en.wikipedia.org/wiki/Rich_Communication_Services>), or aren't supported on all devices (Signal would be excellent, but isn't supported on most Feature Phone / Dumbphone OSes, such as KaiOS/AOSP). And Signal too is ultimately a single provider.
<https://www.t-mobile.com/benefits/scam-shield>
The free tier seems to identify but not block likely scam calls. T-Mobile's website skews sales-heavy, and doesn't feature much technical information.
The premium features appear to be app-dependent, and may not work on feature phones.
For spam mitigations to work, the cost of selling that business must exceed its revenue.
Some, and I won't mention AT&T by name, are very curiously opposed to any regulations touching this.
The other element is that carriers can act at the network level, noting how much abusive traffic arrives from given peers, and taking direct action against those peers. That could involve rejecting traffic outright, subjecting it to stronger challenges, and/or diverting it to investigative / law-enforcement bodies (I'd suggest both national and state entities) for both tracking and enforcement. Power-law relations mean that at any given time, a small number of networks will account for the overwhelming majority of spam, though which networks will likely change over time.
The key problem with this is getting the carriers to act, which ... will probably involve a few carrots and sticks. I'll address those in another comment, except to mention bonding: <https://oag.ca.gov/consumers/general/telreg>.[1]
Individual action will not solve this problem, but there are steps you can take.
Most major US carriers now offer some form of robocall blocking. "Scam Shield" from T-Mobile, "ActiveArmor" from AT&T, "Call Filter" from Verizon.
MVNOs (mobile virtual network operators) may or may not offer scam / robocall blocking themselves (though IMO they should, and should be required to). Some will identify spam calls, but those are still passed through to your handset.
Beyond this, there are on-device apps which can be used, some are carrier-based (e.g., "Call Filter Plus", from Verizon, similar tools exist for Verizon and AT&T), some are third-party. These of necessity share your voice/text activity with third parties, which is its own concern and consideration.
Full Android, iOS, and several full-featured Android alternatives (GrapheneOS, /e/OS, LineageOS, etc.) offer unknown caller rejection. Numbers not in your contact list are directed to voicemail. At present, few spam calls will leave voicemail, though some do, and as AI expands in capabilities, applications, and adoption this will all but certainly increase. I'd strongly encourage use of this.
Feature phones / dumbphones ... have far less capability. Most cannot even reject unknown numbers, which ... seems a ripe target for legislation and/or regulation. Phone frameworks such as AOSP / KaiOS seem to afford little capability for even creating a call-blocking app. This and other dumb devices (e.g., traditional landlines) are a strong argument for carrier/network level mitigations.
The company everyone loves to hate, Comcast/Xfinity, actually has one of the most sophisticated voice/text spam blocking systems, and one I'd like to see mandated to all carriers: <https://www.xfinity.com/support/articles/spam-blocker-overvi...>
It's risk based.
It classifies calls into three categories: high, medium, and low risk.
It adjudicates calls based on risk.
High-risk calls are terminated entirely.
Medium-risk calls are directed to voicemail.
Low-risk calls are subjected to an audio CAPTCHA (enter a two digit value to ring through), otherwise are directed to voicemail.
(It's not clear whether or not a whitelisted number will escape any treatment, perhaps subject to conditions such as originating from the appropriate/approved network for that call.)
I haven't used that system, but in advising people still moving off landlines, or looking at VOIP solutions, it's making Comcast an attractive option.
(I don't know what other VOIP providers, say, Twillo or Asterisk, offer, but suspect at least some have similar if not more-capable systems.)
________________________________
Notes:
1. California requires a $100,000 bond by all telemarketers in the state. The state has a small fraction of the incidence of robocalls of the worst US states. Several others have some bond. My view is that bonding should apply at the carrier level and be surrenderable to both contacted individuals and downstream peering networks, to provide both a strong financial penalty to abusers, and an incentive to downstream networks to pursue abusive calls.
Full disclosure I'm the founder, and I've got a couple dogs in this fight
Broadband Breakfast does seem to be one of those entities. ATIS (<https://atis.org/>) is another, though as a telco alliance I consider it highly suss.
(Submitter.)
Initially the approach was a long list of rules, usually a whitelist and blacklist of known good and bad contacts, and then a large set of specific patterns and assigned weights. Procmail was an early standard here, later Spamassassin.
The biggest revolution came with Bayesian classification. YCombinator's Paul Graham (@pg) developed one such system. The idea here was that a small set of mail was classified into two categories, spam (unwanted) and ham (wanted), and the classifier went looking for patterns within each corpus, automatically assigning weights. This took much of the guesswork and assumptions out of the process, but still relied on contextual clues within the mail itself, though both data (the message payload) and metadata (email headers) could be used.
Following that were reputation-based systems, generally looking at domains or IP address space, where a sufficiently large-scale survey of mail patterns, initially based on honeypots, later largely conducted by large email providers themselves such as AOL, Yahoo, Hotmail, (this was the aughts, they still existed), and eventually Gmail and a few others. Senderbase/Ironport (later bought by Cisco) were another major contender here. These approaches strongly leveraged power-law relations, in which a small number of origins (IPs, CIDR blocks, ASNs) account for the vast majority of email spam. Generally: poor network hygiene practices, whether intentional or otherwise, show, and are actionable by peers / others.
Google especially, through Gmail, had access to a phenomenal amount of activity, and could detect both datacenter-based bulk mailing activity and residential proxy campaigns. Effectively its Gmail service serves as a huge, distributed, collection observatory, and can respond to new spam campaigns incredibly quickly. I don't have specific insights, but suspect that response times are measured in minutes if not seconds.
Google of course also has insight to the contents of emails, but network- and header-level adjudication is much faster, cheaper, and surprisingly effective.
This is why I'm strongly advocating carrier-based, network-level phone-spam mitigations, and whatever regulatory changes are necessary to incentivise providers to adopt these. On-device apps are fine, so far as they go, but would best work in concert with network-level countermeasures.
'Kay. So you call people and just clearly have a tone the second you hear a voice which could well be someone's answering machine
(Of course most of it would be automated, I guess)
Special place in hell for these people
There's not much publicly-available research on the topic. One better source I've found is "Robocalls: A Worldwide or US-only Problem? Analyzing Spam and Fraud in International Phone Calls" (30 Jun 2026)
<https://arxiv.org/html/2606.31790>
On the magnitude in the US vs. elsewhere:
Even though robocalls are an international problem, our findings indicate that US citizens are substantially more affected than the rest of the world. The median number of phone numbers in our US honeypot was 11 751, compared to 101 913 international phone numbers. Despite an almost tenfold difference in the number of available phone lines, the US honeypot numbers received 8 314 813 calls (95.1%) compared to 432 538 calls (4.9%) received by the international numbers. On average, a single phone number in our US honeypot received 707.5 calls during a nine-month period, about 2.62 calls per day. Non-US numbers received, on average 4.24 calls during the same period, about 0.016 calls per day.
The paper does not break out calls per person by country, though it discusses general patterns.
Hiya has an accessible report which highlights "seven key countries", with calls per month in parenthesis: The US (22), UK (5), Canada (6), Spain (13), France (18), Germany (3), and Brazil (29, highest in the world).
Americas: Brazil (29), Mexico (26), Chile (23), and the US (22) lead.
In Asia: HK (23), Indonesia (16), the Philippines (12), and India (8) lead.
<https://work.hiya.com/hubfs/2025/Global%20Call%20Threat%20Re...>
Cheap messaging is as annoying now as calling. Ruining the phone experience overall. Soon Apple will have to do something.
Commerce and crime go hand-in-hand.
The Greek and Roman gods of travel and communication were also the gods of tricksters (frauds) and thieves. The etymology of "Mercury" may be related to that of "merchant". (Hermes rather less so.)
Instant messaging solved all this shit 700 years ago!
Let each number act as an "account" on the phone network/company, just like IM accounts, each with its list of contacts, and blocked numbers.
Shove incoming calls into "Strangers" with low-annoyance notifications (ringtones) by default unless they're from a "verified" company etc.
Oh and yeah as others said, a "Report Abuse" button.
Phone numbers are a scare resource and the telecommunications networks heavily regulated with numerous central points of control.
This bullshit is scaling because the companies which gate and sell that access have no obligations, legal or otherwise, to deny scammers access to their resources.
2-5% of scams succeed but penalty should be there accounting for 95-98% that did not succeed.
IANAL but I know failed attempt at robbery or murder is also prosecuted, failed attempts at scam are not, because people just hang up and move on. Then the reality is society doesn’t have enough resources to deal with "scam attempts" - well we have to focus on murder attempts an plain robbery
And why would you impose a highly-intrusive personal-tracking system across billions of subscribers?
And you know every other country has full KYC for phone connections, right?
What activities are you engaged in which make you think you'd likely be considered a spammer? What specific harms do you see occurring? Might these be related to your present line of business / profession / employment?
What activities are you engaged in which make you think you'd likely be considered a spammer? What specific harms do you see occurring? Might these be related to your present line of business / profession / employment?
SHAKEN/STIR identifies whether or not a given number is originating from a specific network, but without knowing whether that's an approved network, rejecting unapproved-origin calls, or tracking how much unapproved traffic a given network is emitting and penalising it for this, the information isn't actionable.
"Measure it harder" doesn't solve problems. The information must direct meaningful action.
<https://news.ycombinator.com/item?id=49130932>
There's a related issue apparently of small operators who can't afford (or haven't been bothered) to implement STIR/SHAKEN. Many of these are apparently small rural phone co-ops (its own interesting bit of telecoms history). There should be both support in providing them with such capabilities, and penalties for failing to do so, including liability for transiting spam calls to their own or other carriers' subscribers.
Where there's legitimate business, and the risks are low (including the overwhelming majority of personal / residential lines, as well as most business / institutional lines), there's absolutely no problem.
The key is getting incentives right. Indexing bonding and payouts to the spam level, and having a target, is one way to approach this.
The problem with cheap comms, or cheap anything, is that it makes low-return, high-volume activities viable, including especially those which externalise costs and internalise benefits. Such as, say, fraud or spam generally.
What my proposal does is internalise those costs to spammer and telcos which facilitate them, by raising effort (most bulk calls are blocked) and shifting the financial incentives. From participating in the fight against spam for the past 30 years I've a pretty good notion that this will work. And unlike email, the phone system already has a costing and payment system built in.
With single spam operations generating hundreds of millions to billions of calls monthly and barely achieving profitability, I'm pretty sure we can knock things back a lot without inconveniencing legitimate players.
Please walk through as well, what the privacy, surveillance, data-disclosure, third-party bad-actor, and authoritarian-government risks of such a system would be.
And I think this is also a problem at the habit/behavior level for people. Most people don't want to know how to set proper boundaries with people and technology and articulate what they want. Once you do that at least you can articulate what you want to come in at you or not.
Without that you get stuck with a weird one-size-fits-all policy which definitely doesn't fit for me at least.