One thing that probably could happen from that list in 2022 is weaponized ai driven malware.
Вринципе и сейчас запускаются они, но выхлоп конечно же меньше чем был в самом начале....но тоже деньги
This is kinda interesting thing, what kind of AI/ML usage do you think can be implemented in malware and how would it improve current state of malware in general?
I belive that generational AI (both supervised and unsupervised) could be used to reduce detections by only letting the best of the current generation be used as a starting point for the next generation.
Yes, it is possible, but you need to be able to generate a lot of samples from some kind of template and scan them quickly. One interesting approach is to use GANs (generative adversarial network): one neural network learn to detect malware and another neural network learn to break the detections made with the first network. There was kinda good presentaion about it, maybe you and other people here will like it (I had hard time googling it out for some reason):
Well, yes, it is possible to do, but I think it is better to do it using a deterministic algorithm (like a dozen of if-else branches) for simplicity and stability. You will need quite a few simulated environments to make your malware learn them, and eventually if your model will make mistakes (if it is not trained enough) you may lose quite a few targets. Right now I don't see real benefits for using AI/ML here, but why not?
In my opinion the one thing that holds back the potential of AI is that most AI models are hyper specialized in only one thing if a general purpose model were to emerge then I believe that AI would be beneficial for use.Right now I don't see real benefits for using AI/ML here, but why not?
