Would Artificial Intelligence (AI) be valuable in finding Targets in Afgh/Pak?
I’m thinking along the lines of Adm Poindexter’s data mining ideas. We get a lot of Keyhole data. We should be able to automatically mine that data to find safehouses and training camps based on road traffic. An expert system can cross-match the overhead viewing (NRO) with Echelon type data collected by NSA. Based on the parallax between these data streams as mined, one could have a top ten list of plausible targets.
At that point overflights are feasible to drop motion sensors a recording radios. These yield data at a much enhanced level of resolution. So it’s like zooming in. You look at the whole two countries first, get your top ten list, drop some data gathering covert equipment, then zoom in for closer analysis of each of the plausible targets.
It should be possible for one or two targets per day to frame an hypothesis about what’s likely to happen next at that site. This gets fed into the machine.
The machine can then provide an indication if the hypotheses are verified by new incoming data. If that happens, hopefully one has a Predator in the air, locked and loaded, not too far away. Vector it in the the suspect site. Now take your final look through the on-board camera. If the prospective target still looks high value and low collateral damage, put fire and steel on it. Break things. Kill people.
Our main problem is that we don’t have the savoir faire for high quality targeting. In those regions any targeting that relies on local populations to supply HUMINT is going to be low quality targeting. The human sources just tell us which rival tribesmen the source would like to see dead. We always get sent on somebody else’s mission.
If our mission is the vaporize safe houses for Al Qaida and Taliban, and to vaporize training camps of Al Qaida and Taliban, then we need to find a targeting method that is HUMINT-free — not contaminated by duplicitous mendacious deceptive data, but pure, based on Articifical Intelligence and Expert Systems and the concept of multi-layer zooming in, and the concept of hypothesis framing/hypothesis verification (or not).
With better targeting systems we could have a tenfold greater effect on our true mission while pulling 30,000 soldiers out of Afghanistan instead of sending 30,000 more in.
By fighting an Over the Horizon War from Behind the Wire we can never win a military victory of conquering Afghanistan (which we do not need to do), or achieve nation-building (which we do not need to do), but we could vaporize all safehouses and training camps which is the one and only thing that we actually need to do.
It has been said that the War in Afgh/Pak is a "necessary" war. But for what is it "necessary"? Is it for getting the top execs at the major defense contractors their $20 million per year compensation packages? Is that what it’s necessary for? Well, if so, OK then, they can still "earn" those $20 M/year compensation packages, but let their work be slightly different, more life sparing for American soldiers, more cost saving for the US taxpayers, more intelligently done, with newer and better methods, not just the old Viet Nam methods, like in Hue City, walk around until somebody shoots at you, then you can shoot back, or maybe drive around until the villagers get lucky with their IED. These are the ways war must be waged if you want to take and hold ground (we don’t need that), or if you want to win hearts and minds (we don’t need that). Counter-insurgency is a 40 year old strategy that has never worked. Iraq is still up for grabs, Lebanon and Somalia were clear failures, Viet Nam a disaster. Behind the Wire is a brand new strategy that is very likely to work. Doing the same old things and expecting something different to happen is not excellent leadership.
So, would AI be valuable in finding targets? What do you think?
For okiknowki:
No, not anybody carrying an AK. You see, that’s not good targeting. What I’m talking about is building a behavioral model, framing an hypothesis about what one might expect to see if the cross-hairs are on a safe-house for Al Qaida or a terrorist training camp, and, if that hypothesis is verified firmly by the incoming data — yes I want to assassinate, kill, and obliterate without benefit of trial in New York City. I want fire and steel on the target. We are in a war. They are the enemy. The General Charter of the UN has exceptions dealing with nations whose security is at risk. They can take measures as required without violating the Charter. I like your word. The SEAL team I used to do staff work for as their lawyer liked it too. War does involve killing people. And if they don’t get a notification ahead of time and are taken completely by surprize, I guess you could call that assassination – what I want to do to Bin Laden and his pals.The U.N. gonna whine me down?
sure……
first of all, the lack of good intelligence on Alquaeda and the fact we dont know where Bin Laden is only show that these are not top priorities…..but having contractors in Iraq and Afghanistan and spending all that money that they earn is of course a top priority. Besides you do know that if we actually got Bin Laden that would actually hurt profits …right?
2) we can withdraw combat forces except for small elite force located in surrrounding countries that would surveil and launch commando raids when necessary and of course launch bombing missions against specific targets when necessary. In other words, we dont need $200,0000 soldiers there….what we need in probably $10,000 soldiers in surrounding countries who make hit and run raids and bombing missions.