Would it be possible to give a program a immune system to stop hackers?
Yes, it would, and that’s what we call firewalls and antivirus software
I mean ones that adapt to the hacker.
Building an AI into a firewall program may be possible in the future, but it will be difficult work because…
- Hackers use a large variety of different techniques to break into a system
- The firewall would have to be able to avoid seeing its own code or the code of critical programs as malicious (Panda AV recently saw an update that caused it to see the Windows updater or even itself as malicious and bricked people’s machines as a result).
- Hackers always adapt, an intelligent firewall will still need to rely on human programmers and the research of the software company because it might be something beyond the limits of its program
- ‘True’ AI that would actually write new code for the program is still a ways off. Even IBM’s Watson, which is about as advanced as we’ve gotten in AI, works by fetching information already written from the internet.
Also, in terms of preventing the hacking of an entire system, the company behind the Bitdefender software recently came out with a server box that is essentially an instant firewall/antivirus for every machine connected to it so you don’t have to maintain a separate installment of it for every machine in the home or at the office.
there is
SELinux
but the REAL PROBLEM !!! is this
PEBKAC
there is no software that can fix the person
( just read the tech support blowing off steam on dumb users posts )
/r/talesfromtechsupport/
Adaptation would first require recognition. To implement such a system, one would have to be able to anticipate different forms of threats. This is done to some degree with AV software.
On a more cheerful note, deep learning systems might fit the bill in the future. Stronger forms of AI might be able to quickly learn from threats and immunize accordingly. 
Go read Tyro:
http://qntm.org/tyro
Yes it’s fiction, but it’s a good read.
Truth is, any ‘real’ AI requires understanding. For instance, you understand that a virus is intended to harm your computer, and it could do it by erasing the hard drive, changing bits on the hard drive maliciously, flashing messages on the screen etc. But try and tell a computer this. What makes a virus that makes pop-up adds on your screen different to a program that pops up weather information? One you want, one you don’t. The only way to tell if it’s a virus is to understand what the user doesn’t want to happen.
We are an absolute mile off computers doing anything like understanding.
True ‘Deep’ ai exists, and can be trained to learn from others, like Watson,
If Watson ‘Observed’ enough anti hacking activity, he could predict and compensate for digital attacks,
however his abilities are still governed by what he knows,
most hacks, exploit things we don’t know about…
Memristors probably will bring about strong ai.Then computers could have an immune system to stop hackers.Computers are faster than hackers so they should be able to do it.