Will algorithms be better combat leaders?

2015-06-22T23:44:00-04:00

One of the key things computers are better at is taking large amounts of information, processing it, and coming to accurate conclusions. An article from War On the Rocks about swarms, drones, and the future of warfare (linked below) got me thinking--the author focuses on drones doing all the fighting with humans in control centers, each person controlling swarms of drones, while I thought of computer decision making augmenting human soldiers in combat. Would interconnected devices spread across drones and individual soldiers with a distributed computing algorithm be better at coordinating soldiers than humans are? I'm thinking of things like a series of sound sensors on each soldier picking up on gunfire and determining where the good guys and bad guys are based off the sounds, plus tracking when magazines are being changed by whom, how much combat power each side has, and then automatically coordinating teams to win the fight? Could the platoon, company and even battalion become obsolete as computers provide better guidance for operations and logistics?

http://warontherocks.com/2015/02/between-a-roomba-and-a-terminator-what-is-autonomy/

Between a Roomba and a Terminator: What is Autonomy? - War on the Rocks

Editor’s note: This is the first article in a six-part series, The Coming Swarm, on military robotics and automation as a part of the joint War on the Rock

Response by COL Charles Williams made Jun 22 at 2015 11:48 PM

2015-06-22T23:48:53-04:00

Seriously? Did you see Alien? Or (late add) 2001 Space Odyssey and the HAL 9000? I think the human dimension leaders (you) bring are what makes us different from other countries and machines. I can't imagine an Army or weapon without a "man in the loop" but I am also now retired...

Response by Capt Seid Waddell made Jun 23 at 2015 1:00 AM

2015-06-23T01:00:44-04:00

I think automation will vastly increase SA.

Response by Lt Col Private RallyPoint Member made Jun 23 at 2015 1:35 AM

2015-06-23T01:35:29-04:00

We need to educate our personnel to use technological advancements as enhancers to their decision-making and not replace it. It takes losing the "gee whiz" response and asking what these new capabilities really give us.

Response by LTC Private RallyPoint Member made Jun 23 at 2015 1:44 AM

2015-06-23T01:44:03-04:00

Technology is a tool. Nothing replaces human judgement. Also technology could be vulnerable to cyber-attacks.

Response by Sgt Aaron Kennedy, MS made Jun 23 at 2015 6:53 AM

2015-06-23T06:53:47-04:00

I'm reminded of the book Friday by R. Heinlein. In it, there is a pilot (minor character) and he talks about his inability to actually do anything if something goes wrong. The main character comments "then why do they have you if everything is automated for the most part [sic]." His response is both chilling and pertinent:

"Because I'd try. I'd do everything I could thing of." The limitations of computers is that they don't have free thought. They can only do what they have been programmed to do. They don't "learn," they may "absorb" information, but they don't learn.

Using a troop as an example. If we run out of ammo, we switch to a bayonet, then to hands, then to anything available. We're "hardwired" for survival, and we will use anything as a tool whether that item was intended for that purpose or not.

As an extreme example, no computer would have ever been able to anticipate using an aircraft as a missile, because no programmer (up until that point) would have thought to do it.

Now, going to the article, it specifically speaks to autonomy, and how much function is farmed out from the human operator. That isn't necessarily a bad thing. Humans are great at doing a variety of things, but we're really good at doing a SINGLE thing at a time. If the goal is to allow "focus" where the commander is able to remove extraneous data that he doesn't need that moment (and his staff can do the same below him), then an (semi) autonomous system is great.

Response by MAJ Robert (Bob) Petrarca made Jun 23 at 2015 8:17 AM

2015-06-23T08:17:55-04:00

Until the "Terminator" scenario when they become self aware and turn on us. Interesting article.

Response by Capt Richard I P. made Jun 23 at 2015 10:54 AM

2015-06-23T10:54:23-04:00

Eventually synthetic intelligences will replace biologics in all things. The progress of technology is inevitable.

The near term is a drone that doesn't need a man in the loop to make kill choices (because of the vulnerabilities of EM or cyber etc.). Ultimately machine intelligences are capable of more rapid computations, and can indeed learn. If our nation doesn't do it first, someone else will.

Most thinkers on this topic limit themselves by taking a short term and a retroactive view of history, where a more forward looking view is very illuminating. A cloud of micro-drones in a swarm of 500m^3 with 1-4 drones every cubic meter could neutralize an aircraft carrier.

Response by TSgt Private RallyPoint Member made Jun 23 at 2015 2:42 PM

2015-06-23T14:42:56-04:00

I feel like I can weigh in on this, being a Software Engineer of many years. First, semantics; An algorithm is a procedure or formula for solving a problem. An algorithm's job is to solve a specific problem, and should be built with the objective of meeting certain performance criteria. When it comes time to solve a logistical problem, or perhaps calculate the optimal spread of Combat Air Patrols based on statistical analysis of hostile activity, an algorithm may be your best friend. Algorithms tend to work best when tackling large problems like this, as most possible outcomes will be predictable within a certain margin of error. When you view the issues of an individual or a small team, an algorithm is probably not going to be the most useful. While these are very useful pieces of information, there are still countless situations where an algorithm is just not going to be able to understand the issues they face.

I would estimate that algorithm use may be able to streamline the chain of command, but it won't eliminate it. It would allow for faster flow of information in order to inform the decision makers, and allow for consistent results. Of course, that is assuming the algorithms are well designed and engineered in the first place.

Response by MAJ Ken Landgren made Jun 23 at 2015 5:37 PM

2015-06-23T17:37:21-04:00

One of the big challenges is the fact the enemy is adaptable and the battlefield is not always linear. I am sure we will try limited applications and learn from there. You cant program an infinite number of scenarios in the computer.

Response by CW5 Private RallyPoint Member made Jun 24 at 2015 7:33 PM

2015-06-24T19:33:39-04:00

Very interesting topic, CPT(P) Private RallyPoint Member. I think it's great that machines (computers) can process information faster and better than humans, but I would never want to take the human out of the equation. As long as humans are ultimately in control, I say use the technology to the max ... Just don't ever give up the decision making to machines. Here's a TED talk on the (or at least a related) topic:

http://www.ted.com/talks/daniel_suarez_the_kill_decision_shouldn_t_belong_to_a_robot?language=en

The kill decision shouldn't belong to a robot

As a novelist, Daniel Suarez spins dystopian tales of the future. But on the TEDGlobal stage, he talks us through a real-life scenario we all need to know more about: the rise of autonomous robotic weapons of war. Advanced drones, automated weapons and AI-powered intelligence-gathering tools, he suggests, could take the decision to make war out of the hands of humans.

Response by MAJ Ken Landgren made Dec 15 at 2015 3:24 PM

2015-12-15T15:24:36-05:00

Absolutely not. War is full of unexpected and unintended consequences. Battles can shift, terrain changes, threats change, and units who are on the defense can often go on the offense.