21

AI is vulnerable from two security perspectives the way I see it: The classic method of exploiting outright programmatic errors to achieve some sort of code execution on the machine that is running the AI or to extract data. Trickery through the equivalent of AI optical illusions for the particular form of data that the system is designed to deal with. ...


12

Asimov's laws are not strong enough to be used in practice. Strength isn't even a consideration, when considering that since they're written in English words would first have to be interpreted subjectively to have any meaning at all. You can find a good discussion of this here. To transcribe an excerpt: How do you define these things? How do you define "...


7

Programmer vs Programmer It's a "infinity war": Programmers vs Programmers. All thing can be hackable. Prevention is linked to the level of knowledge of the professional in charge of security and programmers in application security. eg There are several ways to identify a user trying to mess up the metrics generated by Sentiment Analysis, but there are ...


6

How we can prevent it? There are several works about AI verification. Automatic verifiers can prove the robustness properties of neural networks. It means that if the input X of the NN is perturbed not more that on a given limit ε (in some metric, e.g. L2), then the NN gives the same answer on it. Such verifiers are done by: Stanford: https://arxiv.org/...


5

The reason this is hard is because it is not trivial to understand what a law means. Many humans still have a hard time understanding laws and thus we have millions of judges and lawyers who study years to be able to even debate whether a law was broken at all. More generally to AI, the problem of understanding laws is a byproduct of the bigger problem that ...


5

There is the case of the tesla accident where the car was in autopilot and crashed into a truck because it appears the vehicle mistook a lightly coloured truck for the sky, killing the driver: https://www.newscientist.com/article/2095740-tesla-driver-dies-in-first-fatal-autonomous-car-crash-in-us/ Having said that, it appears the car had been trying to tell ...


5

https://www.technologyreview.com/s/530276/hidden-obstacles-for-googles-self-driving-cars/ Google’s cars can detect and respond to stop signs that aren't on its map, a feature that was introduced to deal with temporary signs used at construction sites. But in a complex situation like at an unmapped four-way stop the car might fall back to slow, extra ...


5

It's not just Hawking, you hear variations on this refrain from a lot of people. And given that they're mostly very smart, well educated, well informed people (Elon Musk is another, for example), it probably shouldn't be dismissed out of hand. Anyway, the basic idea seems to be this: If we create "real" artificial intelligence, at some point, it will be ...


5

To answer your question, it really depends on the purpose of the Artificial Intelligence program. For example, 4Chan has hacked a number of "Artificial Intelligent" bots, most notably was Microsoft's Twitter bot Tay. The general purpose of the bot was to parse what was tweeted at it and respond in kind, learning and evolving with each and every interaction. ...


4

However, do industrial strength, production ready defensive strategies and approaches exist? Are there known examples of applied adversarial-resistant networks for one or more specific types (e.g. for small perturbation limits)? I think it's difficult to tell whether or not there are any industrial strength defenses out there (which I assume would mean that ...


4

Is Artificial Intelligence Vulnerable to Hacking? Invert your question for a moment and think: What would make AI at less of a risk of hacking compared to any other kind of software? At the end of the day, software is software and there will always be bugs and security issues. AIs are at risk to all the problems non-AI software is at risk to, being AI ...


4

Because he did not yet know how far away current AI is... Working in an media AI lab, I get this question a lot. But really... we are still a long way from this. The robots still do everything we detailledly describe them to do. Instead of seeing the robot as intelligent, I would look to the human programmer for where the creativity really happens.


4

Tesla's technology is assistive, as Alexey points out, so this is not a case of an autonomous system (e.g. an AGI) doing some fatal stunt (the product name AutoPilot is famously misleading). Now on why the car assistance led to this tragic accident, there is some information related to AI technologies. Warning: I cannot find again the source critical to the ...


4

I believe it is, no system is safe, however I am not sure if I can still say this after 20-30 years of AI development/evolution. Anyways, there are articles that showed humans fooling AI (Computer Vision). https://www.theverge.com/2018/1/3/16844842/ai-computer-vision-trick-adversarial-patches-google https://spectrum.ieee.org/cars-that-think/transportation/...


4

As Andrew Ng said, worrying about such threat from AI is like worrying about of overpopulation on Mars. It is science fiction. That being said, given the rise of (much weaker) robots and other (semi-)autonomous agents, the fields of the law and ethics are increasingly incorporating them, e.g. see Roboethics.


4

The infamous Flash Crash of 2010 may qualify. It didn't involve Artificial General Intelligence (which is still a hypothetical) or even "strong narrow AI" (such as AlphaGo) but does involve algorithmic decision-making, which is a form of basic Artificial Intelligence. Algorithmic trading already represents a significant percentage of all market activity, ...


3

It's important to note that ultimately, the statistical methods we currently use in ML research are just that: statistical methods. So when they show some "bad behaviour" it's not because of problems with the statistical methods, but with the data we give it. But if the data we give it are as "genuine and unfiltered" as it gets, then it probably shows ...


3

If a machine learning based AI is "sufficiently smart enough" to be able lie then there is nothing preventing it from lying. This does not mean it can't be persuaded from lying. So just make the AI simple enough to not be able to lie. The reasoning here is that in order for a system to be able to lie, a system must be able to recognize an incentive to ...


3

As far as I know, Tesla cars autopilot is not a 100% AI pilot, it's an assitant: as it detects hands off wheel it slows down, so it's incorrect to speak about AI mistake: it is not trained/designed to drive a car all by itself. A human driver is responsible in that incident.


3

To put it simply in layman terms, what are the possible threats from AI? Currently, there are no threat. The threat comes if humans create a so-called ultraintelligent machine, a machine that can surpass all intellectual activities by any human. This would be the last invention man would need to do, since this machine is better in inventing machines than ...


3

There are a number of long resources to answer this sort of question: consider Stuart Armstrong's book Smarter Than Us, Nick Bostrom's book Superintelligence, which grew out of this edge.org answer, Tim Urban's explanation, or Michael Cohen's explanation. But here's my (somewhat shorter) answer: intelligence is all about decision-making, and we don't have ...


3

A robust ML model is one that captures patterns that generalize well in the face of the kinds of small changes that humans expect to see in the real world. A robust model is one that generalizes well from a training set to a test or validation set, but the term also gets used to refer to models that generalize well to, e.g. changes in the lighting of a ...


2

He says this because it can happen. If something becomes smarter than us, why would it continue to serve us? The worst case scenario is that it takes over all manufacturing processes and consumes all matter to convert it into material capable of computation, extending outward infinitely until all matter is consumed. We know that AI is dangerous but it doesn'...


1

Perhaps what you are looking for is the notion of adversarial attacks on machine learning systems?


1

Everything can be hacked. The solutions found by artificial intelligence can be much more efficient than human solutions, but they can also be confused because of the diversity and immensity of details that our mind possesses. Artificial Intelligence models bring us more secure solutions, but nothing is 100% safe when we talk about information security. ...


1

It seems impossible to prevent that. If someone can make a safe AI from scratch in the near future, then someone else can make a dangerous AI from scratch as well. If all that's needed is a computer (or eventually a robot) it will be really hard to stop people from creating one. Banning computers? Maybe it could prevent it, but that comes with quite a few ...


1

This is an old area of AI called "Plan Recognition", which has about 3.5 million results in Google Scholar. A lot of the modern work is done with classical search techniques coupled with expert domain knowledge, or related reasoning concepts like Hierarchical Task Networks. I'm not aware of or able to find recent research using deep neural networks for ...


1

Another point of view - In safety-critical real world systems, this attack should be evaluated from other aspects as well. In many systems the attack is somewhat mitigated to physical attacks only - for example, you can't add digital noise to a camera used for autonomous driving - you need to print an adversarial e.g. stop sign and locate it in a place, ...


1

I concur with Akio that no system is completely safe, but the take away is AI systems are less prone to attacks when comparing with the old systems because of the ability to constantly improve. As time passes by more people will get in the field bringing new ideas and hardware will be improving so that they are "strong AI."


1

You may be interested in the utility functions of deception: From the abstract of Why Animals Lie: How Dishonesty and Belief can Coexist in a Signaling System. (NIH, 2006)" We develop and apply a simple model for animal communication in which signalers can use a nontrivial frequency of deception without causing listeners to completely lose belief. This ...


Only top voted, non community-wiki answers of a minimum length are eligible