Hacker News Reader

Comment by caleb-allen | Hacker News Reader

A small anecdote.

A few years ago in an undergrad business class, we were having some discussion and the topic of encryption came up during one of my presentations. A student asked a question related to the ethics of encryption (I don't recall exactly what), and I was clearly confused by the question.

To clear up confusion, the professor asked those who thought encryption was "bad" to raise their hand, and at least 60% of the class raised their hands.

It was pretty jarring to me, and makes me pessimistic about the outcome of a DOJ campaign to demonize and regulate encryption

View on HN

I think we in the tech community tend to vastly under-estimate the threat of legal restrictions on encryption. When the public gets scared, they look to governments to "do something", whether that something is really a smart thing or not.

If we're unlucky and we get caught unprepared, we run the risk of getting stuck with a backdoor or "exceptional access" mechanism that provides little or no technical safeguards against massive and nearly unlimited government overreach.

IMO this makes it our responsibility to figure out how we might design such a system that does have strong protections against misuse. Of course that is a very difficult thing to do. Doesn't mean we shouldn't try.

I think you'll run into a chicken-and-egg problem here. Without bonafide strong encryption, those strong protections against misuse probably can't exist.

> Without bonafide strong encryption, those strong protections against misuse probably can't exist.

Agreed. I actually did some work on this topic that requires having things like strong ciphers and hash functions, and protocols that provide perfect forward secrecy.

Our approach was to take strong constructions and turn them into a kind of proof-of-work. So you can recover the key, but to do so you have to expend a huge amount of electricity (ie, money).

The idea was partly just to call the bluff of the government types who claim that "exceptional access" is so critical to our safety. If they're not willing to spend the money, then we should not be willing to compromise our security.

NOTE: I'm not suggesting that we should actually deploy this kind of thing, as long as we have a choice in the matter. Just that we should be prepared in case we need it.

Anyway, the paper is here:

C.V. Wright and M. Varia. Crypto Crumple Zones: Enabling Limited Access without Mass Surveillance. In IEEE EuroS&P, 2018. https://web.cecs.pdx.edu/~cvwright/papers/crumplezones.pdf

It sounds like a majority of the students had no idea what encryption was and because the authority figure (the professor) asked them whether or not it was bad they just went with it? I'm having trouble understanding why people would say mathematical functions are bad.

> why people would say mathematical functions are bad.

People don't think 'encryption' is a mathematical function. They think it's like a radar detector in your car. "Why would you have it unless you intend on speeding?" [goes their thought process]

"Oh I use this frequently and things I do every day would be totally broken without it? Really?!" [is the realization we hope dialogue could bring]

Then again, given speeding is so widely accepted, radar detectors are legal to use in 49 out of 50 states. It would be nice if encryption was viewed the same way.

Isn't (relatively) strong encryption legal in every one of your 50 states and deployed to more than 10^10 devices globally?

No, we absolutely don't want strong encryption viewed as "illegal but accepted", we want it viewed as "absolute necessity for day-to-day life and personal safety, unacceptable to ever weaken".

I think it comes down to popular vote versus government policy. While things like speeding and marijuana are illegal, popular votes have made these things legal (or quasi legal) in their respective jurisdictions. For example, votes to remove speed cameras or make them illegal, or votes to decriminalize marijuana.

But you are correct in that we really should make it completely legal so that the government doesn't have an excuse to target individuals for engaging in a practice that is generally widely accepted.

> I'm having trouble understanding why people would say mathematical functions are bad.

That's such a naïve way of putting it. People don't care about the mathematics behind it. They think it's "bad" because they think it enables terrorism and the spread of child sexual abuse imagery or whatever politicians have led them to believe.

Let's be clear - encryption can absolutely enable those things.

That is not all it does, but pretending it does not is stupid.

The internet enables all of those things too. Before that, telephones and fax machines did.

Hospitals enable the spread of disease by clustering all the sick together with a constant flow of healthy people (visitors, doctors, etc). They also enable the development of superbugs.

What's your point? You're confusing incidental usage with enablement. The lack of these good things doesn't prevent the occurrence of the evils listed, and to make a conversation one sided for how much it exacerbates the evil without instead focusing on the good and the reason they exist is incredibly disingenuous (which is the clearly the intent of most of the people on a soapbox about the evils)

> I'm having trouble understanding why people would say mathematical functions are bad.

Encryption is more than just mathematical functions, it's the application of mathematical functions.

And the application is what matters, because it would be easy to say a Nuclear weapon is just physics or that a bio-weapon is just chemistry.

Clearly, how you use basic principles matters.

It isn't like Nuclear weapons, though, because every single computer has everything you need to do full encryption. Nuclear weapons require materials, manufacturing, etc.

Nuclear weapons require both the knowledge and the physical world parts.... encryption just requires the knowledge, which is impossible to put back in the box.

Every house has the basic materials for a bomb, and many houses have the basic materials for a big bomb (especially if you use fertilizer). The difference is understanding how to use it.

Every computer is capable of building software that can bring society to its knees, but most people don't know how to do it. Software is just algorithms, after all.

The problem is that law enforcement pushes the bad uses of something without weighing it with the good. If fertilizer can be used to make bombs, should be ban fertilizer? No, that's throwing the baby out with the bathwater. Unfortunately, many voters don't fully understand what the baby is WRT encryption, they just see child porn, human trafficking, etc.

Encryption allows data to be locked away from the government including law enforcement and prosecutors in a way that was nearly impossible for the average citizen a few decades ago. Warrants can't break encryption like they could doors or locks. As much of life moves to the digital world and becomes encrypted, that can be a drastic change in how the justice system works. Pro-encryption people need to keep this in mind when talking to anti-encryption folks. Criminals being able to hide evidence is a serious concern for the average person.

> in a way that was nearly impossible for the average citizen a few decades ago

Ehh, maybe if you are talking about speed and scale. The average citizen could always learn how to use a one-time pad.

Speed, scale, and by default with no initial sharing of keys are super important distinctions.

Why can't the government serve warrants to give up their private keys?

I don't really see the State's argument against encryption except in the case that they want to break encrypted information without due process.

Well, 5th amendment protections, and inability to determine whether a person even remembers the key, means that serving a warrant doesn't always work either.

But then, neither did breaking down the door and ransacking the obvious physical target pre-internet days always work. And everyone understood a suspect had no obligation to tell the feds where he hid the goods.

I always have trouble imagining examples of this kind of evidence that criminals were previously utterly unable to hide but can now do it easily using encryption. It seems any kind of evidence could had just been shredded and burned in times past or hidden in an obscure, physical safe that no one knows about. If anything, the digital age had made the location of evidence more obvious and even produced entirely new kinds of evidence (traffic and connection logs, emails, etc).

It seems to me that law enforcement is simply pining for the fjords of a very short time span in human history when ample digital information co-existed with the lack of widely available cryptographic systems.

> Criminals being able to hide evidence is a serious concern for the average person.

Suspects. Alleged criminals. Until convicted, they're not criminals. The government does not have an unlimited right to collect all information; they have a limited, judicially controlled right to try to collect information.

One key detail: strong encryption does not prevent investigations from targeted collection of information, such as through physical surveillance, bugs, etc. There are many tools available to law enforcement, those tools are just less convenient. Framing it that way helps: in order to make investigations more convenient, the DOJ wants to prevent anyone from using secure communications.

>Suspects. Alleged criminals. Until convicted, they're not criminals.

And even after they are convicted, they and their associates are still protected by encryption. Encryption protects everyone equally. That is viewed as a positive by pro-encryption people and a negative by anti-encryption people.

>The government does not have an unlimited right to collect all information; they have a limited, judicially controlled right to try to collect information.

This the point of warrants. The problem of massive warrantless collection of data by the government is a bigger and separate issue. I personally think encryption is a valuable tool to address the symptom of that problem, but it does nothing to actually fix problem of a corrupt and/or authoritarian government. One of my personal annoyances with the tech community is that we have a tendency to try to work around any governmental problems as if that solves the issue while making no attempt at fixing the root cause.

> One of my personal annoyances with the tech community is that we have a tendency to try to work around any governmental problems as if that solves the issue while making no attempt at fixing the root cause.

Agreed. We need to produce technical solutions, and effect policy solutions, and get a lot better at PR and rhetoric.

The only way your argument makes any sense is if you can somehow show that conviction rates have gone down since encryption became widespread. I think you'll have a hard time doing that because people will usually find some other way to hide information, like just destroying evidence. Are you gonna go after shredders next? They're hiding criminal's secrets too!

Warrants might not be able to break encryption, but refusing to provide the passcode can land you in jail indefinitely for contempt. You could serve decades potentially.

The fear from law enforcement is absurd. They've never had the ability to read every single letter that was sent, but now that we've moved online they want to read every email? They claim they are losing access but the truth was they never had access in the first place. The only difference is before maybe they could secretly read your letters before delivering them but now they need to get your password. That only worked at a small scale though, and I'm sure the NSA can crack passwords at those same small scales (e.g. some of the high-profile iphone cracking we've seen).

> Encryption allows data to be locked away from the government including law enforcement and prosecutors in a way that was nearly impossible for the average citizen a few decades ago.

I know some people do think this, but the opposite is actually true, because of the popular shift from voice to computerized text and images for almost all communications.

A few decades ago, criminals spoke on the phone to coordinate actions. If the police were aware of this in advance, they could tap the phone line and listen in to that conversation. However if they found out about it after the fact, there was no record of the conversation for law enforcement to go get. The sands of time hide information extremely well.

Today, we use mobile devices, and prefer text and image-based communications--email, SMS, WhatsApp, iMessage, Facebook, Twitter, etc. etc. Unlike voice conversations of the past, these default to creating a permanent record of the conversation, which police can go get after the fact. Encryption is an attempt to re-establish the privacy and security of these conversations that existed back when they would have been a voice call over analog telephony.

The idea of "going dark"--promoted by law enforcement to fight encryption--is wrong. What has actually happened is that INCREDIBLE BRIGHTNESS of information has developed over the past few decades as all our computerized communications now default to living forever.

Encryption is a tool for trying re-establish the level of "light" (information availability) that our culture and laws came to expect over hundreds of years.

I hope those people access their bank over port 80.

They do, on their open wifi network.

I'm having trouble understanding why people would say mathematical functions are bad.

Because the vast majority of people don't know that encryption is mathematical formulas. They only know what they see from Hollywood, which is that it's a thing that bad guys use.

Should have followed up with: Do you think having a lock on your front door is bad?

Eh. Locks can be picked and doors can be rammed. Layered encryption is really hard to break and one time pads are impossible.

So locks are only good because they don't actually work? And encryption is only good if it doesn't work too?

That logic kind of makes sense. However, locks at least delay intruders and make it harder for someone to go undetected if the break into your house. The intruder is either at your door for a little while, makes noise trying to break in, etc.

If the police bust into my house there will probably be witnesses and I will know about it. If they steal my data, how is anyone going to know? This gives them way too much power.

They already have a great deal of power. Technological progress gives us all more power.

I think locks are a bad analogy. Going online without encryption is like living in a house without walls or windowshades. It's less about physical access and more about people being able to see what you're doing. Even if you're not doing anything "wrong", the lack of privacy still feels uncomfortable.

So in an ethics discussion the professor simply asked "is X bad?" That doesn't seem like a very enlightened ethics discussion.

How certain are you he said "bad"? If he has instead asked "is encryption problematic?", the outcome can be interpreted much differently because there are problems with encryption. Especially in an ethics discussion, there are definitely pros and cons to encryption. (FWIW, the pros outweigh the cons by far, IMHO.)

I'm not trying to say your memory is wrong, I'm just asking how certain you are about what was asked exactly. Changing one synonym for another can make a big difference.

Per GP this was in a business class likely taught by a business professor, not an ethicist or someone similarly specializing in such discussions. It's not surprising that they'd pose the question to the group in such a blunt fashion. As a CS major, I had to take an engineering and a CS ethics course and it was similarly problematic. The courses were taught by an engineering or CS professor (respectively) and ethics discussions were not their specialty.

Asking if it's 'bad' is an excellent starting point for an ethics discussion. It reveals pre-existing biases and emotional responses to a topic, which must be taken into account as a pre-requisite to any rational discussion. You cannot get dispassionate without first examining the passions.

Asking about a nuanced subject in broad and blunt terms can be pedagogically useful because you hear what people really think, unfiltered ... and maybe they hear it too, priming them to realize that they've been underthinking the topic. From there you can get into nuance, which is easier when you know where you currently stand.

I have no idea what tack this prof took with the question, but I think you've drawn a hasty conclusion here.

I didn't intend to conclude anything. I was just asking if he remembers exactly what the professor asked, and saying that the exact phrasing of the question changes my interpretation of the results. I do not expect him to remember, nor do I think there is anything wrong with his comment, but I wanted to ask.

>That doesn't seem like a very enlightened ethics discussion.

That is an apt summary of my college ethics discussions. Even in cases where the professor was doing their best to encourage discussion, the majority of the class couldn't think beyond first order effects and hypotheticals would quickly get emotional and devolve into character attacks.

Yeah. I was shocked at how many -CS MAJORS- were pro-DRM.

What does being a CS major having to do with being pro-DRM? Isn't it mainly an economic/philosophical standpoint on copyright?

I believe it's something more to do with the OP's (mistaken) idea that CS majors are somehow more likely to be "hackers" or people who exchange pirated software or wares -- when really there's only a handful of people who know how to crack the DRM in the first place (which requires some degree of technical know-how) and then those people distribute the pirated goods to others (whose technical literacy may be far less, even if it's still enough to distribute the pirated material on down the line).

It's a bit like thinking all painters are in favor of graffiti, or all musicians are okay with "sampling" and remix culture.

Actually, it comes from the author's belief that CS majors should know that, A. DRM at best delays cracking, never prevents it, and B. That it ends up hurting the legitimate users, not the pirates, and C. That the more efficient it is at A, the more likely it will cause B. As even when I was in college, that was the case (I asked some of the students to give me an instance where a piece of desirable software was not available illegally within a week of release date. They were unable to provide one (in fairness, since then there have been a few cases of less desirable releases that went longer; nothing that had the attention of all the major cracking groups). I also asked if they were aware of some of the issues at the time (such as securom's huge issues, the installation of rootkits by some, the inability to use when moving computers, etc), they were not, despite having explicitly chosen the topic to do research on for the class. They went into the topic with the expectation that DRM is just a good thing, and even when researching it, only sought out sources that enforced that bias. Which shocked me, given they should be technically literate enough to know the issues, and the greater nuances involved.

An interesting follow question should have been how many people think privacy is bad. That leads to a discussion of encryption as a form of privacy boundary...

What was the class about? What sort of majors were the attendees?

Maybe the professor should have asked whether drugs are bad. (and yet this is still a poor analogy)

Similar but different anecdote. I installed i2p the other day to see what had changed/improved over the last year, and mentioned to my 12yo son that I had done this. He asked what i2p was, and I said it was the dark web. He immediately thought he knew what this is, and asked "aren't you worried about getting in trouble because it's illegal?"

Reply on news.ycombinator.com