A few months ago an essay titled “Cybersecurity is not very important” appeared. The essay is well written and interesting but I’d like to argue against its main point.
And that is actually hard – the essay has many good points, and although it has a contrarian feel, it actually isn’t saying anything outrageous. But I still don’t agree with the conclusion. I suggest reading it (or skimming it) first before continuing here, although this article is generally self-sufficient.
I agree with many things in the essay, most importantly that there is no 100% protection and it’s all about minimizing the risk. I also agree that cybersecurity is a complex set of measures that span not only the digital world, but he physical one as well. And I agree that even though after watching a few videos from DEF CON, BlackHat or CCC, one feels that everything is fundamentally broken and going to live in the mountains is the only sane strategy to survive an impending digital apocalypse, this is not the case – we have a somewhat okayish level of protection for the more important parts of the digital world. Certainly exploitable, but not trivially so.
There are, though, a few main claims that I’d like to address:
- There has not been any catastrophic cybersecurity event – the author claims that the fact that there was no digital Pearl Harbor or 9/11 suggests that we’ve been investing just the right amount of effort in cybersecurity. I don’t think that’s a fair comparison. Catastrophic events like that cost human lives as an immediate result of a physical action. No digital event can cause immediate loss of human life. However, it can cause indirect loss of human life, and it probably has already – take a famous data breach in an extramarital affair dating site – do we know how much people were killed in Pakistan or Saudi Arabia because infidelity (or homosexuality) was exposed? How many people died because hospitals were victims of ransomware? How many people died when the Ukranian power grid was attacked, leaving 20% of of Kyiv without power and therefore without heat, light or emergency care? What about the (luckily unsuccessful) attempt to sabotage a Saudi Arabia petro-chemical plant and cause an explosion? There are many more of these events, and they are already a reality. There are no visible explosions yet, which would make it easier to compare them to Pearl Harbor or 9/11, but they are serious and deadly nonetheless. And while natural disasters, road incidents and other issues claim more victims, there isn’t a trivial way to calculate the “return of life on investment”. And isn’t a secure charity for improving hurricane protection in third world nations better than one that gets hacked and all of its funds get stolen?
- People have not adopted easy security measures because they were minor inconveniences – for example 2-factor authentication has been around for ages, but only recently we began using it. That is true, of course, but the reason for that might not be that it has been mostly fine to not have 2FA so far, but that society hasn’t yet realized the risks. Humans are very bad at intuitively judging risk, especially when they don’t have enough information. Now that we have more information, we are slightly better at estimating that, yes, adding a second factor is important for some systems. Security measures get adopted when we realize the risk, not only when there is more of it. Another reason people have not adopted cybersecurity measures is that they don’t know about them. Because the area is relatively recent, expertise is rare. This discrepancy between the ubiquity of information technology and the lacks of technical expertise (not to mention security expertise) has been an issue for a long time.
- The digital world plays too small a role in our world when we put things in perspective – humans play a small role in the world if you put them in a big enough perspective, that doesn’t mean we are not important. And the digital world is playing an increasingly important role in our world – we can’t that easily continue to claim that cybersecurity is not important. And overall, the claim that so far everything has been (almost) smooth sailing can’t be transformed into the argument that it is going to be the same, only with gradual improvement over time. If IT is playing an exponentially more important role (and it is), then our focus on information security can’t grow linearly. I know you can’t plot these things on a graph without looking stupid, but you get the gist.
- We have managed to muddle through without too much focus on cybersecurity – yes, we have. But we will find it increasingly harder to do so. Also, we have successfully muddled through many eras of human history because we have done things wrong (For example the Maya civilization collapsed partly because they handled the the environment wrong). Generally, the fact that something hasn’t gone terribly wrong is a bad argument that we are doing fine. Systemic issues get even more entrenched while on the surface it may look like we are successfully muddling through. I’m not saying that is certainly the case for cybersecurity, but it might very well be.
While arguing with the author’s point is an interesting task, it doesn’t directly prove the point that cybersecurity is indeed important.
First, we don’t have good comparisons of estimates of the cost – to the economy and to human life – of investment in cybersecurity as opposed to other areas, so I don’t think we can claim cybersecurity is not important. There are, for example, estimates of the cost of a data breach, and it averages several million dollars. If you directly and indirectly lose several million dollars with a likelihood of 30% (according to multiple reports), I guess you should invest a few hundred thousands.
Second, it is harder to internalize the risk of incidents in the digital world compared to those in the physical world. While generally bad at evaluating risk, I think the indirection that the digital world brings, contributes negatively to our ability to make risk-based decisions. The complexity of the software complicates things even further – even technical people can’t always imagine the whole complexity of the systems they are working with. So we may not feel cybersecurity is important even though facts and figures show otherwise.
But for me the most important reason for the importance of cybersecurity is that we are currently laying a shaky foundation for our future world. Legacy software, legacy protocols and legacy standards are extremely hard to get rid of once they are ubiquitous. And if they are insecure by design, because they are not built with security in mind, there is no way that software that relies on them can be secure.
If we don’t get cybersecurity right soon, everything that relies on the foundations that we build today will be broken. And no, you can’t simply replace your current set of systems with new, more secure ones. Organizations are stuck with old systems not because they don’t want to get new and better ones, but because it’s hard to do that – it involves migration, user training, making sure all edge cases are covered, informing customers, etc. Protocols and standards are even hard to change – see how long it took for TLS 1.3 to come along, for example. But network standards still have vulnerabilities that don’t have good mitigation (or didn’t have until recently) – take an SS7 attack on a mobile network, or ARP spoofing, or BGP hijacking.
If we don’t agree that cybersecurity is very important, future technology will be based on an insecure layer that it will try to fix with clumsy abstractions. And then at some point everything may collapse, at a moment when we are so dependent on it, that the collapse will be a major disruption in he way humanity operates. That may sound futuristic, but with technology you have no option but to be futuristic. We must build systems today that will withstand the test of time. And this is already very hard – maybe because we didn’t think cybersecurity is important enough.
I’m not saying we should pour millions into cybersecurity starting tomorrow. But I’d be happy to see a security mindset in everyone that works with technology as well as in everyone that takes decisions that involve technology. Not paranoid, but security conscious. Not “100% secure or bust”, but taking all known protection measures.
Cybersecurity is important. And it will be even more important in he upcoming decades.