Note: The article is also on authors medium account.
Here are some questions that came out from current war between Hamas and Israel!
How to detect terror attacks?
How to find terrorists and their hiding places?
How to take militants out so that the loss of common people’s life is minimal?
How to make such AI?
How would such AI work?
How long shall it take to make such AI?
The recent attack on Israel by militants and terrorist group Hamas, makes it more clear to be able to detect terrorism much before time, and hence such attacks may be reduced.
Terror crimes done by Hamas were too bad and completely unpredictable. How can all this happen when there are many kinds of intelligence on terror activities? This is truly a work which has gone unchecked. Hence, from this, we learn that such terrorists may need to be identified using advanced Artificial Intelligence. It should be understood that some work may already be in use by intelligence units across the globe, if it does not use AI this is the time to use terror detection using AI, and if not then time to make, check, and install such intelligence.
The question is how the attack went unchecked, and why intelligence could not detect it. Maybe because Hamas works from an underground caves and tunnels? How and when the tunnels got made, why its neighbor Israel did not know about the tunnels? Who supplied Hamas with the network capacities to launch this scale of attack?
Why can’t there be intelligence now to detect where the militants are? Can AI help in this? If yes then this may save innocent people’s lives, as terrorists attacked Israel not common people.
Well, there can be several ways to see this problem. One can be using AI to detect words that are threatening over suspected calls, other can be to detect keywords related to terrorism activities. Many steps can involve multi-level inference.
Why we don’t have open hackathons and challenges for terror activity detection? Yes, this may lead to some privacy issues. So how else terrorism such as recent Hamas attacks can be detected on time, and found within the timeframe and finally resolved?
Solution is detecting the network terrorist use, satellite data or messages that can be dangerous in calls. This does not require AI, but AI can complement independent decision making before, after or on the plan to detect such things.
The following AI based techniques can be used:
- Knowledge-based decision making
- Uncertainty based solutions
- Deep Learning based solutions
- Search using heuristics
- Game Playing based solutions
- Inference engines
- Classification based on algorithms such as Naive Bayes and so on…
These would be explained in coming times.
Would you allow some data from your side to be allowed to train for terrorism detection?
The only problem in using AI would be data use, this may lead to privacy issues. It is for the common men and women to answer —
Would you mind if some suspicious messages may be read by AI algorithm? Else how else AI could help us and civilizations?
Would you allow your routers to be scrutinized once in a while? To detect unintended use by terror-related agents?
All these with guarantee to not to misuse data and respect for privacy of data.
How long all this may take? As soon as some data is allowed, sand settles over privacy issues and general public wants it, that is all that is needed.
To be continued……..