Posted on Nov 4, 2015
1LT William Clardy
14.7K
104
64
8
8
0
Apparently, the Army is attempting to use "big data" data-mining techniques in an attempt to identify who is most likely to become violent.

Is anybody here troubled about the potential adverse impact this could have on a soldiers' career, especially if the prediction is based upon spotty data analysis?

https://gcn.com/Blogs/Emerging-Tech/2015/11/Big-data-violence-prediction.aspx?s=gcntech_041115&p=1
Avatar feed
Responses: 22
CSM Michael J. Uhlig
8
8
0
Same thing as when we return from down range.....we know how to answer the Doc so we can get out of the gym and go have a cold one.....
(8)
Comment
(0)
SSG Audwin Scott
SSG Audwin Scott
9 y
Good one CSM!
(1)
Reply
(0)
SSG Todd Halverson
SSG Todd Halverson
9 y
So very true.... say the true answer and you end up wasting more time there and getting referred to various medical departments. Only answered truthfully after my last deployment, because I new I was dropping my paperwork for retirement. Then, i did not care if I had extra crap to do because I new I would get to enjoy many cold ones at home.
(3)
Reply
(0)
COL Vincent Stoneking
COL Vincent Stoneking
9 y
Very true, not only there, but in a variety of situations where the troops all know the "correct" answers - regardless of the truth.
(4)
Reply
(0)
1LT William Clardy
1LT William Clardy
9 y
Definitely, COL Vincent Stoneking.
Being told that "There is no right answer" is a warning sign we all recognize, one which ranks right up there with "This won't hurt much", "I sent the check", and one other which I'll omit in the name of family-friendliness.
(0)
Reply
(0)
Avatar small
Capt Richard I P.
8
8
0
Aren't we paid to become violent?
(8)
Comment
(0)
1LT William Clardy
1LT William Clardy
9 y
Capt Richard I P., I thought our ability to do violence on demand was a professional necessity, but maybe we're kinder and gentler military now...
(2)
Reply
(0)
SSG Audwin Scott
SSG Audwin Scott
9 y
1LT William Clardy - Very good point, we are after all trained to kill!!
(1)
Reply
(0)
Avatar small
MSG Brad Sand
5
5
0
IF they can predict when soldiers are going to be violent...considering that is what we train them to do...why not try using it trying to figure out who is a threat to this country? Oh, they have tried that and it doesn't work nearly as effectively as we need? Okay, there is your answer.
(5)
Comment
(0)
COL Vincent Stoneking
COL Vincent Stoneking
9 y
1LT William Clardy - I think you just said the same thing I did in my "note"....
(1)
Reply
(0)
1LT William Clardy
1LT William Clardy
9 y
COL Vincent Stoneking, I would like to think that I was agreeing with your note whilst rephrasing it in an earthier tone and avoiding the use of trademarked buzz-phrases. :-)
(1)
Reply
(0)
SFC Cryptologic Network Warfare Specialist
SFC (Join to see)
9 y
Can they predict when Soldiers are going to commit DUI? sexual assault? suicide? If not, why not? 
(0)
Reply
(0)
MSG Brad Sand
MSG Brad Sand
9 y
SFC (Join to see)
They cannot predict if something will or will not happen. They can predict the probability that you may do something. Committing a DUI is even more murky because it would have to factor is drinking levels and time when one is actually driving. There are a whole bunch of moving pieces that makes these questions difficult to predict.
(1)
Reply
(0)
Avatar small
Avatar feed
Does the idea of the Army trying to predict who will become violent scare or reassure you?
SPC Treatment Medic
3
3
0
Scares the shit out of me. Many of us in the military are going to come up as high risk on that. Just walk around Sand Hill. What other profession out there has "in training employees" shouting kill every other step?
(3)
Comment
(0)
Avatar small
Capt Seid Waddell
3
3
0
Sounds like the "Minority Report".
(3)
Comment
(0)
1LT William Clardy
1LT William Clardy
9 y
Summed that up pretty succinctly, Capt Seid Waddell.
(1)
Reply
(0)
Avatar small
CMSgt James Nolan
3
3
0
1LT William Clardy Correct me if I am wrong, but the military needs people capable of violence to serve does it not?
Coming off of my last deployment, we had to stop by the Medical Group for out-processing. So, I am sitting there getting questioned by a 22-23 year old medical technician and she asks me "Do you ever think about hurting anyone?" So I respond "Every day". She is has this look on her face like I just crushed a puppy, and I tell her that I am a First Sergeant, and I think about it all the time. Then, I say, "Was that my outside voice?" The Doc who is at his desk behind her, literally fell out of his chair laughing, she is in disbelief, and I have to assure her that I have done this before, and no, I don't want to hurt anyone....
If the system works as well as all of that, we are doomed LOL.
Seriously though, question: Just because bad things happen, and people come from rough backgrounds, do we really want to call them in to see the Doc, because a program has decided they are more likely to get violent? I am all for people not wigging out and going bananas, but not for a computer to isolate troops based on probability equations.
"Sergeant, the computer says you need to go see the Doc...." Sounds kind of big brother-ish...
(3)
Comment
(0)
Avatar small
CPT Manager
3
3
0
If "we" weren't violent...would we go to war?
(3)
Comment
(0)
Avatar small
CPT Military Police
3
3
0
(3)
Comment
(0)
SSG Audwin Scott
SSG Audwin Scott
9 y
That's exactly my point CPT M Cannonie, it has to be an inaccurate study.
(0)
Reply
(0)
CPT Military Police
CPT (Join to see)
9 y
SSG Audwin Scott I didn't see your comment to the poster in the string of comments, I only responded to the posters comment.
(0)
Reply
(0)
SSG Audwin Scott
SSG Audwin Scott
9 y
CPT (Join to see) - It's alright
(0)
Reply
(0)
1LT William Clardy
1LT William Clardy
9 y
Even more critical, CPT (Join to see), is that they never mention what their false-positive rates have been. The signal-to-noise ratio is crucial -- identifying the most violence-prone 5 percent means very little if the list contains 10 or 15 percent of the evaluated population.

I also was just noticing an intriguing tidbit I overlooked previously: Judging solely by the article, the recidivism rate for violent offenses must have gone up if more than half the violent crimes by soldiers since 2010 were committed by just 5 percent, while the most violence-prone 5 percent in the previous decade only committed something more than a third (but implicitly less than half) of the offenses.
(1)
Reply
(0)
Avatar small
SrA Matthew Knight
2
2
0
Sounds like Operation Insight which was the main bad guy plot point from Captain America: The Winter Soldier. So long as it isn't used to arrest or punish people who haven't even done anything wrong just because the computer thinks they will then it shouldn't be a problem though.
(2)
Comment
(0)
Avatar small
SGM Mikel Dawson
2
2
0
Seems to me this is like predicting the weather.
(2)
Comment
(0)
Avatar small

Join nearly 2 million former and current members of the US military, just like you.

close