4 ms·
Is anybody else worried that a security-vulnerability in a popular family of internet-connected self-driving car could allow a worm that could cause casualties
by aidenn0 10y ago
Is anybody else worried that a security-vulnerability in a popular family of internet-connected self-driving car could allow a worm that could cause casualties on the order of millions? I mean how many people would die if e.g. every late-model Ford in the US started aiming for pedestrians?
- aetherson 10y agoI think that there are a bunch of ways that you could make this very difficult for teenage idiots to manage, and if you're talking well-funded, highly competent state actors, they probably don't want to do terror attacks like this (and if they did, they could probably cause comparable harm today. Always remember how easy it would be to get a few sticks of dynamite into most US trains or subways).
- beatpanda 10y agoRemember that time when a well-funded, highly competent state actor stockpiled vulnerabilities when everyone told them not to and the leaked vulnerabilities nearly brought down an entire national health care system?
- wfunction 10y ago> they could probably cause comparable harm today. Always remember how easy it would be to get a few sticks of dynamite into most US trains or subways That's localized death on the order of thousands at most, not global death of millions like the parent post was suggesting. Still obviously horrifying, but doesn't really compare in terms of the threat.
- aetherson 10y agoWell, one bomb would be, yes. A sufficiently hypothetically scary driverless car attack might manage to do more deaths (though it'd be really hard, even for a state actor). But I think it wouldn't really get to the level of "more capable of causing terror." If you could cause let's say 15,000 deaths by a well-coordinated one-day bombing attack, and 80,000 deaths by a well-coordinated one-day driverless car attack, do those two things actually differentiate themselves in terms of your nation's agenda? I'd think they'd cause more-or-less identical amounts of terror, and the only purpose of such attacks are terror. Ultimately, I don't think it's important. Nobody who's resourced well enough to pull this off is interested in doing it.
- aidenn0 10y agoIf I understand correctly, your claim is that only well-resourced (e.g. state-level) attackers will be able to pull this off. It is extremely clear to me that today both desktop computers and IoT devices do not require a well-resourced attacker to compromise (can we agree on this?). The point where we disagree is that I see little evidence that driverless car companies are investing in defense in depth that is at a level significantly higher than either IoT devices or desktop operating systems. Indeed, there is such a race to be first-to-market that the companies are incentivized to not spend resources on it.
- monk_e_boy 10y agoIt's going to make stealing cars easy. Use a 0-day, log in, carry on picking up passengers, route the money to me.