The Vigilantes Who Created 'Malware' to Secure 10,000 Routers
forbes.com
forbes.com
For a full answer, I also need to answer some other questions. Is there any way to contact the owner of the device that has a reasonable chance of success? Are there any reasons why the owner might leave the device in an unsecured state intentionally? Is the device spreading harm to others because it is insecure? Is the owner aware that the device is not secure?
I decided that the scenario was analogous not just to someone leaving their car doors unlocked in a high-crime neighborhood, but to a medical quarantine. An unlocked car doesn't really hurt anyone else, but an unaware carrier of a contagious pathogen is dangerous to himself and to others. It then becomes more acceptable to force treatment on an unwilling or unaware person, for the sake of public health.
As a result, I figured that a program like the following would be ethically acceptable, but not necessarily virtuous:
- uses a single specific vulnerability to compromise systems
- immediately patches the vulnerability out of an infected system
- attempts to detect and remove any known malware
- goes dormant for a variable amount of time
- attempts to infect other systems at a slow rate for a fixed amount of time
- removes itself from the infected system
- leaves behind a human-readable log of its activity
It would also be helpful for it to be accompanied by a fix for the vulnerability in the official device update stream.The medical quarantine example is interesting, but there is a critical difference: There are officially recognized medical authorities with the legal power to require you to be quarantined. Currently, no such authorities exist on the network.
So the example is more like: some people with medical knowledge but no actual standing try to quarantine you. But it's not quite as bad as it sounds, because there exists no Health Department to do so officially. (For that matter, there doesn't exist much government at all, in any form.) What's more, since there is no government to enforce this, the self-appointed expert quarantines you without your permission.
I can see the reasoning behind it, but I'm not willing to go so far as to call it moral/ethical behavior.
It's very easy to violate someone's self determination rights with the best intentions and still hurt them. In the virus parallel the vector is an unknown quantity. It doesn't entirely seal the router, and doing it without permission may in fact increase the vulnerability of the router (or brick it, if things go wrong).
I find this really interesting, but it's very hard to back from a moral or ethical point of view.
The end result benefits me slightly, yet the means does not entirely justify it.
If I lived in the tropics, I'm not certain how I would feel if I returned home to find that a band of rogues had broken into my house, installed new insecticidal bed nets in all bedrooms, and locked up after themselves on their way out, without touching anything else. On the one hand, hey, free bed nets. On the other hand, no one bothered to ask my permission. Or knock.
But then maybe I hear about a terrifying disease currently spreading via the saliva of the oogeyboogeyfly, who will chew through untreated bednets, or insecticidal nets older than 6 months, to bite sleeping people, and only new ones treated with skellingtol will work to stop them.
In that case, there was
- an overriding concern for public safety, from an actual threat
- some urgency, such that voluntary compliance might be too slow
- a risk that some might not take effective preventative measures
- no malicious intent toward me, specifically
The target of the action was actually the oogeyboogeyfly, which I was unknowingly providing with a safe platform for spreading disease. So it comes down to the old Star Trek 3 dilemma: if you could save a thousand lives by letting just one person die, would it be ethical to do it? Would it still be ethical if that one person wasn't you?If you intentionally get yourself infected with a zoonotic pathogen and start cavorting with one of the carrier species, at minimum, you're going to get a severe beating when the town finds out about it. If you don't take care of yourself as a person, someone else might take care of you as their problem.
My opinion is that forced vaccination is acceptable when everyone receives uniform compensation for their loss of self-determination, with prepaid care for all side effects and complications. That means whomever is doing the forcing has to weigh the benefits of universal vaccination against the massive additional cost for compensation and insurance.
So you're not going to force vaccinations for rhinovirus. You might for smallpox, or polio, or ebola, or something else that costs a hell of a lot to control by other means.
And, in the actual situation we're talking about, those doing it did not take on the cost of compensation and insurance.
The concept of authority is simply a synonym for overwhelming force in my ethical reasoning, wherein any reasonable person would flee it or surrender to it rather than openly resist it. So when you say duly constituted authority that just tells me that the teams of doctors have enough men with guns behind them that they don't have to bother with explaining their actions or chasing down stragglers.
In the situation under discussion, the cost of compensation would certainly be little more than pocket change, especially given that the victims may not have even noticed the intrusion, or even the more invasive prior compromises. And we don't know the actual cost of damages to the victims.
The grey-hats' failing was in not allocating resources for those things, however small they may be, before launching the program. Did they sit down and ask themselves, "What will we do for the people who don't think we're doing the right thing?" and "How will we fix anything we might screw up with this?" If not, they probably should have, but now that the deed is done, I won't get worked up over it.
There are VERY FEW ARM based routers. Huge majority runs MIPS. You know what runs ARM? stupid IoT devices, security systems/cameras, and custom/very specialized niche jobs.
Here is a graph from the original story: http://www.symantec.com/connect/sites/default/files/users/us...
The risk with these things is always the "what if?". What if during that patch it bricked my device due to some fluke? What if I had a special customization that relied on a specific behavior on the component that was just updated without my consent?
What if that patch attempt to a vehicle fails, leaving it unable to turn on? What if that vehicle was an ambulance?
And to flip that argument again - would you prefer your device to be bricked by someone trying to run a webserver serving child pornography, or by someone trying to protect you from the first?
If you patch routers and 1% get bricked by some fluke, is that not a win for society?
Also, to rebut the "How dare you change the firmware on my router without my consent?!" point, let's not forget these were essentially unprotected routers at this point in the applied_patch vs known_vulnerability cycle. If you don't want anyone patching your router without your permission... SECURE YOUR OWN PROPERTY.
If I leave something on the sidewalk, I don't get angry when people cart it off / step on it. I especially don't get angry when they glue the broken bits back together and leave it on my doorstep with a note saying "We noticed this was broken. Superglue works wonders! Here's the number for a hardware store where you can buy it."
If the vehicle is an ambulance, a well-intentioned intruder could break it at any time, whereas a malicious one could break it at the worst possible time. While either one could cause a disaster while carrying a trauma patient at full speed on the highway, the accidental damage might also occur while motionless and empty in a parking lot.
If grey-hats can patch a vulnerability against your will, it is very likely that they first tried disclosing the vulnerability to the manufacturer of your device, and they declined to fix it within a reasonable time, or you didn't run that update. Either way, whatever damage done to the device as a result of exploiting vulnerability would not have occurred if the vulnerability had already been repaired.
If a router ships with a brickable vulnerability, the warranty SHOULD cover it. I don't know if it WOULD.
Analogies just over-abstract the issue.
But there has always been a tension between the utilitarian criteria you're using, and the absolutist* criteria that others use when evaluating the morality of a decision. In practice, the most effective systems seem to strike a balance. Something like: "First, do no harm. Everything else is details. And the details matter, and the details need to be regularly debated, and the debate itself is an important part of the process."
Abstract examples like the trolley problem don't teach us things that we can't conceive of. They can help us boil down and isolate specific aspects of absolutist thinking and determine if we're better off with the statistical approach to deciding if an action is good or bad.
* Technically referred to as deontological ethics, but dang, software engineers ain't got nothing on philosophers when it comes to oblique terminology.
This is still quite a big violation of privacy, even if was done for good reasons.
It's like some one notices that your window is open so they pick the lock on you front door and enter your home to close it.
Beautifully put. And they justify doing it because of the danger of someone breaking and entering. But someone might have broken and entered, and they definitely did, so the justification sounds very hollow.