APCO P25 PHASE 2 ENCRYPTION
- Thread starter bocook222
- Start date
- Status
And why is that?
They claim GPT4 can do so much it only seems like a matter of time before it infiltrates encryption since it is supposed to be able to write computer code. Encryption is just one of many things A.I. Will create issues both good and bad. We'll just have to keep up with the news as it is being discussed on a daily basis.
I am not worried about this for P25 radios. Assured voice communications is what my users need....and I can just have users change the key if needed. Unless GPT4 can assign unique radio ID's in a P25 system, or somehow takes over Provisioning Manager, I am not worried.
lwvmobile
DSD-FME
Oh boy, let me put my tin foil hat on for this one....they're in the walls.
AI is being way overrated! Not to mention extensive, exaggerated fear mongering. 
mmckenna
I ♥ Ø
I think it remains to be seen how AI would guess encryption codes fast enough to do any good. The limits have always been set by computing capacity.
Either way, periodic rolling of encryption codes should stay ahead of that. Consumer scanners don't have the facilities to handle encryption, anyway.
And there will be just as much effort from the opposite direction, using new technology to increase the security of communications.
Either way, periodic rolling of encryption codes should stay ahead of that. Consumer scanners don't have the facilities to handle encryption, anyway.
And there will be just as much effort from the opposite direction, using new technology to increase the security of communications.
avaloncourt
Member
- Joined
- Jun 10, 2011
- Messages
- 167
I agree with mmckenna. The same limitations for AI would exist as they are now. You can't process enough data to stay ahead of it and, as noted, changing up code release would put things right back to the beginning. Unless AI noticed a previously-unrealize flaw, nothing is going to change. Quantum computing on the other hand...
WPXS472
Member
I believe that anything encrypted with AES256 is safe for the time being. When I read that quantum computers had actually become used routinely, I did a little research. It seems that the algorithm used in AES isn't amenable to solution by quantum computers. At least not yet.
From what I was able to determine, in the not too distant future, the three latter agencies will probably have the ability to defeat any of the currently available encryption methods. But, this will probably not be in real time. Things will have to be saved for analysis later. As we all probably know, there is one system which cannot be broken, One time pad. But it is difficult to implement, and not suitable for radio communication systems. As to the threat posed by AI, I think there is a lot of hysteria at present. Sure, there are dangers. But, the actual folks doing AI development are well aware of them. I believe they are responsible enough to use AI as a tool. As it should be used.
From what I was able to determine, in the not too distant future, the three latter agencies will probably have the ability to defeat any of the currently available encryption methods. But, this will probably not be in real time. Things will have to be saved for analysis later. As we all probably know, there is one system which cannot be broken, One time pad. But it is difficult to implement, and not suitable for radio communication systems. As to the threat posed by AI, I think there is a lot of hysteria at present. Sure, there are dangers. But, the actual folks doing AI development are well aware of them. I believe they are responsible enough to use AI as a tool. As it should be used.
mmckenna
I ♥ Ø
The use of AI, if it does ever happen, will be used in both directions.
The systems will not stay at AES256, newer algorithms will come out.
AI may be used by some to try and hack it, but there will also be AI behind encryption keeping things more secure.
Encryption does get periodically updated. Used to be frequency inversion scrambling was "good enough", then DES, now AES. It'll keep rollng along, trying to stay one step ahead. Agencies should also be rolling their keys periodically, so anyone discovering the key shouldn't have much time to enjoy it.
As for AI and those developing it.
I'd like to think they are "responsible enough", but we know that is not the case. Someone is going to see a way to make money, and there won't be any amount of rules/laws that will stop them once they smell the green. It -will- get used and abused in the name of making people richer, and that will come as a detriment to our society.
The systems will not stay at AES256, newer algorithms will come out.
AI may be used by some to try and hack it, but there will also be AI behind encryption keeping things more secure.
Encryption does get periodically updated. Used to be frequency inversion scrambling was "good enough", then DES, now AES. It'll keep rollng along, trying to stay one step ahead. Agencies should also be rolling their keys periodically, so anyone discovering the key shouldn't have much time to enjoy it.
As for AI and those developing it.
I'd like to think they are "responsible enough", but we know that is not the case. Someone is going to see a way to make money, and there won't be any amount of rules/laws that will stop them once they smell the green. It -will- get used and abused in the name of making people richer, and that will come as a detriment to our society.
WPXS472
Member
Quote:
As for AI and those developing it.
I'd like to think they are "responsible enough", but we know that is not the case. Someone is going to see a way to make money, and there won't be any amount of rules/laws that will stop them once they smell the green. It -will- get used and abused in the name of making people richer, and that will come as a detriment to our society.
I understand your concerns, and do share them. This reminds me of the whole genetic engineering thing. There were fears of "designer babies" a few years ago. That has not happened. For one thing, human DNA and how to manipulate genes to achieve a certain result without unexpected consequences has proven very difficult. There was that one instance where a Chinese researcher managed to give AIDS resistance to a child. Or, so he says. He went to jail for it. In general, researchers in that field are being very careful not to run afoul of public opinion. AI researchers might be less cautious. The obvious rewards of coming up with "the next big thing" are enormous. But, there is an awful lot of work involved as well, and most won't want to risk that for fear of adverse public reaction. AI is coming. It remains to be seen whether it will be a blessing or curse.
As for the safety of encrypted systems, it is far easier to bribe, or coerce someone with access to encryption keys to divulge them than it is to try to actually break the encryption itself. People are always the weak link.
As for AI and those developing it.
I'd like to think they are "responsible enough", but we know that is not the case. Someone is going to see a way to make money, and there won't be any amount of rules/laws that will stop them once they smell the green. It -will- get used and abused in the name of making people richer, and that will come as a detriment to our society.
I understand your concerns, and do share them. This reminds me of the whole genetic engineering thing. There were fears of "designer babies" a few years ago. That has not happened. For one thing, human DNA and how to manipulate genes to achieve a certain result without unexpected consequences has proven very difficult. There was that one instance where a Chinese researcher managed to give AIDS resistance to a child. Or, so he says. He went to jail for it. In general, researchers in that field are being very careful not to run afoul of public opinion. AI researchers might be less cautious. The obvious rewards of coming up with "the next big thing" are enormous. But, there is an awful lot of work involved as well, and most won't want to risk that for fear of adverse public reaction. AI is coming. It remains to be seen whether it will be a blessing or curse.
As for the safety of encrypted systems, it is far easier to bribe, or coerce someone with access to encryption keys to divulge them than it is to try to actually break the encryption itself. People are always the weak link.
lwvmobile
DSD-FME
I prompted Stability AI to create some photos of "Florida Man"...yep, pretty sure its just a matter of time before Skynet takes over and cooks us all and the Matrix turns us into human batteries for the machine.
mmckenna
I ♥ Ø
Yes we are.
Most research universities have some level of ethics review. In other words "We CAN do this, but should we?" Unfortunately research grants are a tempting thing, and sometimes the ethics get overruled.
Industry, not so much. If they smell a dollar, it's good to go. Until they get caught.
mmckenna
I ♥ Ø
The polydactylism is a nice touch (no pun intended).
lwvmobile
DSD-FME
Oh, trust me, that wasn't intentional, that is literally how AI draws hands and fingers. That's one of the 'better' ones, at least they are mostly shaped correctly.
mmckenna
I ♥ Ø
avaloncourt
Member
- Joined
- Jun 10, 2011
- Messages
- 167
Really? Didn't you see the recent articles on the Google AI chatbot? Internal people on the project lobbied for it not to be released. They said it was a compulsive liar and often gave responses which could result in human injury or death if instructions were followed. The Google ethics board gave it the green light because Microsoft was releasing theirs.
I wish y’all wouldn’t use big words like that. Fortunately I know how to use Google.
mmckenna
I ♥ Ø
I'll give that can a high-6
All of us are counting base 10, cat be like base 12 (or 14).
All of us are counting base 10, cat be like base 12 (or 14).
- Status
