9
$\begingroup$

In the near future our net (digital) activity is monitored and logged into kind of social credit system. How would our protagonist go about

  • Hide illegal activity (Could it be camouflaged as legal maybe?)
  • Generate acceptable activity. (No activity is suspicious.)

Botting activity is a no go as bots (dumb ai) can be detected by other similar algorithms. Humans have unique irrational pattern to their activity.

$\endgroup$
10
  • 3
    $\begingroup$ Close vote? I'd like feedback included with disapproval. $\endgroup$
    – pinegulf
    Commented Apr 21, 2023 at 5:54
  • 2
    $\begingroup$ typically people vote to close for really open ended brainstorming style questions, instead favoring questions where you've thought a fair amount about the situation and are looking for additional feedback or ideas you might have missed $\endgroup$ Commented Apr 21, 2023 at 14:52
  • 4
    $\begingroup$ "Humans have unique irrational pattern to their activity." you VASTLY overestimate human activity. Here is a fun fact - Google, Amazon, or whatever other service doesn't "listen" to provide you with ads that are relevant. Simply because it doesn't need to. Humans are far more predictable than you and many give them credit to. Harvesting bits and pieces of usage data is enough to build up a profile, link it to similar ones and then provide suggestions based on the cluster you fall into. $\endgroup$
    – VLAZ
    Commented Apr 21, 2023 at 16:39
  • 4
    $\begingroup$ Humans are hella dumb, and are easily fooled by dumb AI, as recent technological advances are clearly showing. But more importantly, what you're asking is incredibly broad. "Illegal activity"? It is challenging to be less specific than that! You should rewrite your question to be about a specific aspect of online anonymity or deniability or whatever. $\endgroup$ Commented Apr 21, 2023 at 18:50
  • 1
    $\begingroup$ @pinegulf I didn't cast the VTC, but it's important to understand that no one is obligated to explain a close vote. It's civilized, polite, courteous, helpful, informative, and kind... but not obligated. I forget at what Reputation you can see the "close" link. If you can't, the VTC was for "Too Story-Based," meaning the caster believes there's too many plot permutations to rationally and objectively answer the question. To a degree, I get their point. You don't explain what the technology can do. You simply say "Social Credit System" and assumed we'd all understand. $\endgroup$
    – JBH
    Commented Apr 21, 2023 at 19:00

6 Answers 6

26
$\begingroup$

The answer is sadly the way that most current illegal activity is done online - Mainly through the use of coded language that has both a commonly understood meaning and an alternate meaning known to those who are engaged in illicit activity.

Take, for example, the recent crusade against Child Exploitation on Twitter. No one accused of Twitter of actually hosting anything illegal, but the organization and communication between parties using language that at a glance would not attract attention:

"We have just come into possession of some 2011 vintage Sauvignon Blanc, from the vineyards of Eastern Europe. Our Sommelier can arrange a private tasting session, but spaces are limited."

Now, that message on its own doesn't raise any suspicion to a lay person and definitely not to a bot. However, someone intimately familiar with Wine might raise an eyebrow at East European Vineyards or what a 2011 vintage is - but again, it's got enough plausible deniability - someone may very well have a small business selling obscure wines.

Or it could be that 2011 vintage describes the age of the victims, Sauvignon Blanc denotes their ethnicity (it's a white wine), and the 'location' of the Vineyard could either be where they are being held or where they were abducted from - Sommelier tasting - well, I'll leave that to your imagination.

The more 'authentic' you can make the posts and the more obscure or esoteric you can make your hidden messages, the more likely it is to slip through any automated system and even a manual review system.

Suppose there's a website and a phone number - you call the number and are greeted by a Wine seller - you might decide 'Oh well, this is legitimate, it's just a bit odd' - How difficult would it before for an organized gang to rent a money laundering front operation that has a legitimate business? The answer is not very difficult.

$\endgroup$
8
  • 2
    $\begingroup$ Is that white wine example based on a real case? I've been wondering how pedos always seem to manage to form rings. $\endgroup$
    – user86462
    Commented Apr 20, 2023 at 9:51
  • 2
    $\begingroup$ @AncientGiantPottedPlant - Not that I know of - I made it up for the purposes of answering the question. $\endgroup$ Commented Apr 20, 2023 at 16:05
  • 1
    $\begingroup$ Note for context: there are actually Bulgarian and Romanian Sauvignon Blanc wines from 2011 on the market: vivino.com/BE/en/daria-estate-sauvignon-blanc/w/…. So a tweet like that would be reasonable, weren't it that Sauvignon Blanc is best drank between 18 months to 2 years after vintage at the latest, and doesn't keep for over a decade. $\endgroup$
    – Nzall
    Commented Apr 21, 2023 at 10:36
  • $\begingroup$ also remember: the more data there is, the smaller percentage gets reviewed by a human. Google does not review everything on YouTube (183 hours uploaded every minute) and that's how ElsaGate happened. Two people communicating in slightly odd normal messages is unlikely to get spotted as long as they don't use any trigger words. Once a Fed spots it once, wine-posting gets inserted to the automated trigger system and the whole scheme unravels very quickly. $\endgroup$ Commented Apr 21, 2023 at 17:32
  • $\begingroup$ @user253751 - Indeed - but normally in that case, they look at the account based activity - e.g. they 'know' one account is dealing in illicit acts, so they look at how and with who they are communicating with etc. etc. Some words are very difficult to trigger a Bot off of, because of their common and widespread usage. $\endgroup$ Commented Apr 21, 2023 at 20:52
9
$\begingroup$

When you say "social credit system," I assume you're talking about something like this:

The Social Credit System (Chinese: 社会信用体系; pinyin: shèhuì xìnyòng tǐxì) is a national credit rating and blacklist being developed by the government of the People's Republic of China. The social credit initiative calls for the establishment of a record system so that businesses, individuals and government institutions can be tracked and evaluated for trustworthiness.

There's a small but distinct difference between trustworthiness and criminality. But it's important. My father is fond of a phrase, "you can stand a politician being a fool, but not a crook." The last thing you want is your criminals looking perfect. That's often as obvious as being a crook. Here in the U.S., we sometimes find illegal migrant workers driving under the speed limit. They think they're avoiding discovery by not doing anything to get pulled over by the police. What they're really doing is putting a big flag on their car that says "look at me!" Believe me, they're noticed.

Real people make mistakes. But there are limits even to this. Make a big enough mistake and you're watched constantly. If your Social Credit System is competent enough, then it would notice when a citizen decided to use a TOR node....

Misdirection

A number of answers have given you specific examples of misdirection. Basically, the idea is to hide in plain sight. Yes, you could use military-grade encryption or create an online persona that's the epitome of the perfect citizen — but things like that get noticed. Misdirection means conducting your business in a way that appears to be any legitimate business.

And that's not uncommon in history. They're called Front Organizations.

A front organization is any entity set up by and controlled by another organization, such as intelligence agencies, organized crime groups, terrorist organizations, secret societies, banned organizations, religious or political groups, advocacy groups, or corporations. Front organizations can act for the parent group without the actions being attributed to the parent group, thereby allowing them to hide certain activities from the authorities or the public.

Front organizations that appear to be independent voluntary associations or charitable organizations are called front groups. In the business world, front organizations such as front companies or shell corporations are used to shield the parent company from legal liability. In international relations, a puppet state is a state which acts as a front (or surrogate) for another state.

The idea here is to bury the criminal activities within the normal operations of a legal, moral, and upstanding business or group.

A specific front organization: identity theft

A serious mistake would be to create a fake online persona. A social credit system will be very capable of discovering an incomplete person, someone with no history, no mistakes, no depth of behavior... just a list of facts (no matter how "complete") that don't express a living, breathing person. This is one reason why a front organization would be useful.

But equally useful would be a stolen identity. Now, to be fair, a major problem on the Internet is that it's getting easier to trace Internet activity. After all, somewhere along the line one computer needs to know enough about another computer to effectively talk to it. But let's ignore that. Finding the identities of people who have similar interests and/or behaviors to the criminal's desired misdirection would be an otherwise effective way to skirt the law. The stolen identity, which is complete from the Social Credit System's perspective, wouldn't be obviously identified as a "person of interest." The original identity's trustworthiness acts as a screen for the criminal's activities until caught — in which case the stolen identity takes the fall.

But there's another possibility: chaos

I can't speak to the effectiveness of online social tracking, but if documentaries like The Social Dilemma are to be believed, then what a truly diabolical Social Credit System would do is act as a two-way street.

  1. Tracking people to assess their social trustworthiness.
  2. Influencing people toward what the government believes is trustworthy behavior (like voting for the right person, dontchaknow).

I'll be honest with you, that second idea combined with the discussion from The Social Dilemma makes me wonder if your online criminals can effectively operate in that kind of world. The behavior of the criminal would need to be influenced consistently and predictably by the Social Credit System or it would be automatically flagged as something suspicious.

But...

Here's where the idea of chaos comes in. Computers are good at dealing with patterned systems. They're not particularly good at extracting patterns from chaos (or, worse, they tend to over-extract patterns from chaos, finding order where none exists). Said another way, your clever criminals could find a way to operate within the noise margin of the system.

A noise margin is the minimum "signal strength" a signal must have, below which it's integrity cannot be trusted due to the presence of uncontrollable noise, such as static from any number of sources. In other words, if your radio broadcast falls below the noise margin, there's too much static to predictably hear the music.

All data mining operations have a noise margin, meaning they always acquire data that has little to no relevance to the intent of the operation. Such information corrupts the value of the data you're actually seeking. Software must be written (as effectively as possible, this can be REALLY HARD) to identify and exclude "noise."

The social credit system will have the same problem. There will always be quirks in the system. Your moral, upstanding citizen gives into temptation and buys a copy of Playboy, an oddity as he's never done that before. Or maybe it's someone with the same name that looks like the moral upstanding citizen. Or maybe the data collection system erroneously attached the moral citizen's ID number to Ducky McNeal, who religiously purchases every copy of Playboy.... Noise.

Your criminals, if they work really hard at it, could conduct their business inside the data noise margin, allowing them to conduct business pretty much without ever being caught.

Please don't ask me how to specifically identify examples in the present Internet world of what I just explained. I know just enough about it to be dangerous, but I don't believe I know enough about it to write a credible book. However, I'm a firm believer that if you write a good story, explaining every detail of how this can be done would be boring and unnecessary for most of your readers. And if you write a bad story, it won't matter how detailed you are.

$\endgroup$
8
  • 1
    $\begingroup$ I was thinking more akin to Big Brother 1984, but more modernized and updated, but this works as jumping off point. $\endgroup$
    – pinegulf
    Commented Apr 21, 2023 at 6:59
  • $\begingroup$ In China, not being trusted by the government is probably the worst crime you can commit... $\endgroup$
    – jwenting
    Commented Apr 21, 2023 at 7:08
  • 1
    $\begingroup$ I wouldn't recommend stolen identities -- apart from plot-hooks, if getting flagged is part of the story. Any good system flags glitches for reviews, as any glitch is a potential bug in disguise. This means that stolen identities may work short-term (very short) but the longer they are in use, the longer the chance glitches will occur and be noticed: how can A buy a magazine in City X and connect from a computer in City Y within 2 minutes? Well, they can't, obviously, so someone needs to investigate which event was mistagged. If you're smart, you don't want investigations... $\endgroup$ Commented Apr 21, 2023 at 7:58
  • $\begingroup$ @MatthieuM. and stolen identities would rapidly lead to worse than simple pollution of the system with flawed data, it's lead to an epidemic of people getting their identities stolen so everyone would end up with a rock bottom credit score. The Party wouldn't mind if it didn't lead to their own bureaucracy also being affected by that, but they will be. $\endgroup$
    – jwenting
    Commented Apr 21, 2023 at 10:01
  • 1
    $\begingroup$ "There will always be quirks in the system. Your moral, upstanding citizen gives into temptation and buys a copy of Playboy, an oddity as he's never done that before." here is an anecdote that can serve as an example of "quirks in the system": I've set up regular donations to Wikipedia. My bank at one point automatically blocked them for being suspicious. That's after I've been sending the donations for about two years (with a previous bank) and it was after four months (with the new bank after they had a merger). Somehow the system automatically decided that this one payment was odd. $\endgroup$
    – VLAZ
    Commented Apr 21, 2023 at 16:52
7
$\begingroup$

The rules of social credit system are bound to have any number of legal loopholes, which can be exploited relatively safely.

In the early 1990s, Polish law defined a van as 'a vehicle in which cargo space is separated from passenger space', and charged less road tax for vans than for passenger cars. As a result, this city car became the country's most popular, umm, van. All you had to do was to put a wire mesh between the back seats and the boot: within weeks, you could buy one made specifically to match the car's internal dimensions in any respectable car parts shop.

If you're like most people and are not planning on doing something bona fide criminal, this is how you circumvent a social credit system. As a bonus, a social credit system cannot penalise sufficiently popular loopholes, because it would undermine itself by doing so. (Think what would happen to the airlines' bottom line if say a quarter of their potential customers found themselves forbidden from flying on planes.)

$\endgroup$
10
  • 2
    $\begingroup$ "As a bonus, a social credit system cannot penalise sufficiently popular loopholes, because it would undermine itself by doing so" tell that to the Chinese government... $\endgroup$
    – jwenting
    Commented Apr 21, 2023 at 7:07
  • 1
    $\begingroup$ @jwenting, thank you. But why would I? It already knows, including from its own experience. (It's one of the reasons why it eventually had to give up on the Cultural Revolution, for instance.) It's just us who don't know the loopholes, and therefore cannot notice how the Chinese government looks the other way; but that tells us more about our own ignorance of the matter than anything else. $\endgroup$
    – ihaveideas
    Commented Apr 21, 2023 at 8:06
  • 4
    $\begingroup$ The current system they use penalises using loopholes, no doubt about it, by lowering the credit score of those who use them when the loopholes are found. The CCP doesn't care about being popular, only about being in control. The cultural revolution didn't fail because it wasn't popular (it was), but because it was TOO successful and led to the complete destruction of China's education system and academia, and with that its capability to produce high tech weaponry and other equipment needed by the State. $\endgroup$
    – jwenting
    Commented Apr 21, 2023 at 9:58
  • 1
    $\begingroup$ @ihaveideas: I'll disagree on this one. The main difference between a totalitarian government and an airline is that the former has a monopoly. People can't escape, propaganda and heavy repression ensure they remain tame enough, etc... Remember that totalitarian governments can (and will) change their stances over time; for example China went from penalizing multiple children to encouraging them once the population explosion was curbed. The people, most of whom do not want any hassle, will simply comply with the new rules. Add in reaction speed (nipping it in the bud), and it won't ever work $\endgroup$ Commented Apr 21, 2023 at 12:40
  • 2
    $\begingroup$ @ihaveideas they know what's expected of them: do what the Party says. That means no loopholes. The Party doesn't suddenly decide to punish everyone who didn't spend last Thursday tying balloon animals, or forgot to wear a prime number of sock layers, but it does punish everyone who converted their car into a van by adding a wire mesh to help their social credit. $\endgroup$ Commented Apr 21, 2023 at 17:54
5
$\begingroup$

Impersonation

There are many variations on the theme 'do not steal too close to your own home'. Digitally this should count just as much. You want to look legitimate yourself, so the digital 'you' never does anything illegal. To do illigal stuff you'll have unrelated digital identities.

How can you do this? A fake might be hard to do, so you take a real person their digital identity. Social engineering is a great way to receive people their username and passwords, but also a copy of their sim card for example. Do this with people completely unrelated to you. If you have enough information you can impersonate them, doing the illegal activities you want. Preferably not from your own home with location services on.

Of course you communicate with code, try to further hide behind VPN or the like, but the true camouflage is simply be or look like someone else. You will make a mistake, or techniques improve that they can track you better. If they track something that is not you at all, you're much safer.

$\endgroup$
5
$\begingroup$

Some use of AI could actually help you here

So let me give you a bit of a frame challenge. AI is becoming far more advanced than your common bot, and can even trick it’s way past CAPTCHAs https://gizmodo.com/gpt4-open-ai-chatbot-task-rabbit-chatgpt-1850227471 AI are likely only going to get better at thwarting these types of systems. AI are also already very good at imitating certain humans (think deep-faking, voice and speech emulation, seem to be able to catch on to your likes, dislikes, and usual activity very easily, etc. Another advantage of AI is that you could practically enlist an army of AIs to also create false identities, create distractions, misdirections, etc. that make it far harder to figure out the source of this crime. So an AI could generate fake internet activity for this guy, while the guy is really doing shady stuff. I also agree with another answer that sadly criminals tend to hide their stuff in plain sight in form of a code. I hope you find this answer helpful.

$\endgroup$
1
  • $\begingroup$ Outsourcing is certainly an option. $\endgroup$
    – pinegulf
    Commented Apr 20, 2023 at 12:34
1
$\begingroup$

Decentralized, anonymized "dark web"-style platforms

"Oh but those would be abolished". No, not really. Remember that the current Tor network was basically made viable in the early days by US military intelligence. The very organizations you would expect to crush such kind of activity. All you need is some incentive to keep an anonymous potentially-illegal network operational. IRL this was the motive of governmental agents needing covertness in foreign theaters and for this needed non-spy activity as "noise" to hide in.

As long as any sort of "foreign nation" or space exists, this incentive is still strong. It might not be the Tor network as we know it, but it's base ideas with newer encryption schemes are likely to hold up well.

Because no matter how good your AI: from the outside, it's just noise. And if necessary, the network could also produce internal actual noise between its nodes. Maybe realtime communication would be impossible due to possible input-output matching, but if messages can be randomly delayed, a demasking of users could easily be prevented.

Again, the only thing required for this is multiple state-level actors with some opposition to each others. A unified global government could simply make accessing such a network illegal (and this would likely be trackeable).

$\endgroup$
2
  • $\begingroup$ Very good. Hmm. how would they keep infiltrators out? -Just thinking aloud- $\endgroup$
    – pinegulf
    Commented Apr 24, 2023 at 10:01
  • $\begingroup$ @pinegulf they don't. Infiltrators are the goal. The noise that your own agents hide behind. Btw this is not ficticious, I'm basically regurgitating the lecture I had about it in peer 2 peer systems. It's value comes from the each nation believing that either they get more value out of it than criminals/enemies OR that banning it is technically too complex/expensive. In addition: In China however it is blocked/banned and only available through further complicated workarounds, so story-wise this network not being around is equally viable. In Russia it is somewhat blocked but not effectively $\endgroup$
    – Hobbamok
    Commented Apr 24, 2023 at 10:40

Not the answer you're looking for? Browse other questions tagged .