Imagine the applications and implications of robotics and artificial intelligence in 2045

Machine Augmentation to Staff Functions

Describe your idea

Deep learning artificial intelligence could augment traditional Army Staff Functions. It could adhere to doctrine and best practices while evaluating and optimizing thousands of courses of action. Logistics, personnel, intelligence, and operations could be fed data from thousands of blue force sensor and ISR platforms to come up with optimal solutions.

Top improvement

1 year ago
jah27 said
0
Open Close
Recent research in decision making has examined hu [...]
Recent research in decision making has examined humans' use of algorithms as a decision aid, which tends to produce superior judgments in many cases. However, emerging research shows that more experienced decision makers often shy away from the use of algorithms as a decision aid; the reasons are not well understood. With this in mind, part of this project should look at the psychological and social implications of leveraging machines to augment staff functions with the aim of identifying barriers to implementation and adaptation.
Add Improvement

What makes this idea new or different?

Solving problems in complex, non-standardized environments under high stress is the domain of human decision making. Until now.

Human-machine decision making hybrid structures will combine the advantages of machines with advantages of humans.

Top improvement

1 year ago
chem said
0
Open Close
Primitive versions of this idea exist today in a v [...]
Primitive versions of this idea exist today in a variety of domains; one example is the sports world, where coaches in the NFL and in MLB make decisions that are informed by quantitative experts in order to optimize outcomes for specific situations (e.g. 2-point conversions, pitching substitutions). Advancements in deep learning, distributed inputs, and natural language processing will increase the domain applicability and analytical power of this approach.

In 2045, platoon leaders could be asking their personal JARVIS to plot routes for avoiding enemy detection in megacity environments.
Add Improvement

What will be the implications of this idea?

Human staff officers will set up the problem and perform quality control. This will allow us to work faster and make better, more informed decision. We will operate within the enemy’s decision making cycles, seizing the initiative permanently.

Top improvement

1 year ago
mjensenwv said
1
Open Close
How much will we trust fallible human staff office [...]
How much will we trust fallible human staff officers with personal and professional biases to quality-check the work produced by a trusted algorithm? What kind of quality control would they be performing? The benefit of an algorithm is that it could crunch huge amounts of data to produce an outcome far faster than a human could. In order to do any meaningful QC on the output, the human would need to crunch that same data to see the rationale behind decisions being made.
Add Improvement

What other players said

1 year ago - redcell254 said
Describe your idea
0
Open Close
Perhaps the greatest potential improvement that co [...]
Perhaps the greatest potential improvement that could result from the use of AI in this field is the calculation of potential courses of action branches and sequels. There was a project several years ago named "Deep Green" that was aimed at this goal, but the state of the art was not up to the level to produce useful results. That state of the art might be sufficient now.
1 year ago - redcell254 said
What will be the implications of this idea?
0
Open Close
We will need to adopt new mechanisms of training t [...]
We will need to adopt new mechanisms of training to enable our staff officers to oversee these processes and understand what they are seeing. Some of the AI capabilities will be so complex and subtle that we might not grasp the second and third order effects until they have played out before us.
1 year ago - shattuck said
What will be the implications of this idea?
0
Open Close
The use of advanced analytics to improve the abili [...]
The use of advanced analytics to improve the ability of staff officers to build products would be of great benefit to the Army. Especially the S2. Having the ability to analyse the ground capabilities of the enemy and having current feedback of blueforce movements would be highly beneficial. This system could be augmented if every Soldier became a sensor with biometeric sensors being added to Soldiers equipment.

However, just having a system that makes great products is useless if it cannot talk to the other systems. There are many great army systems, but there is no common language between the systems. New systems and software is often incompatable with legacy systems software. Additionally, the limits of the AI would be determined by the inputs. Garbage in, Garbage Out. The limits of the AI to help the staff officer will be determined by the questions asked by the staff officer.

Another consideration is the impact of system vulnerability as systems are connected together.
1 year ago - chem said
What makes this idea new or different?
0
Open Close
Primitive versions of this idea exist today in a v [...]
Primitive versions of this idea exist today in a variety of domains; one example is the sports world, where coaches in the NFL and in MLB make decisions that are informed by quantitative experts in order to optimize outcomes for specific situations (e.g. 2-point conversions, pitching substitutions). Advancements in deep learning, distributed inputs, and natural language processing will increase the domain applicability and analytical power of this approach.

In 2045, platoon leaders could be asking their personal JARVIS to plot routes for avoiding enemy detection in megacity environments.
1 year ago - jah27 said
Describe your idea
0
Open Close
Recent research in decision making has examined hu [...]
Recent research in decision making has examined humans' use of algorithms as a decision aid, which tends to produce superior judgments in many cases. However, emerging research shows that more experienced decision makers often shy away from the use of algorithms as a decision aid; the reasons are not well understood. With this in mind, part of this project should look at the psychological and social implications of leveraging machines to augment staff functions with the aim of identifying barriers to implementation and adaptation.
1 year ago - mjensenwv said
0
Open Close
I think the implementation would need to start at [...]
I think the implementation would need to start at lower decision-making levels. 22 year old platoon leaders and 30 year old company commanders are possibly more likely to trust machine-aided decision making than 40 year old battalion commanders. As those platoon leaders and company commanders use more tools, though, they'll gain trust and acceptance at higher levels. When battalion and brigade commanders see junior leaders making more effective decisions, they'll start to spread implementation.
1 year ago - jah27 said
What will be the implications of this idea?
0
Open Close
One implication of this approach is that it develo [...]
One implication of this approach is that it develops the "manned-unmanned teaming" idea with one of the easier tasks and in one of the easiest environments: well-structured cognitive activities in office or office-like environments. The work flows and problem solving structure for many staff functions are pretty well understood and in some cases mapped.

Another implication is that this idea has much greater reach than say, trying to apply AI to front-line combat platforms. Staff-related work drives the activities of organizations of various sizes, some of which are very large - Service staffs, Combatant Command staffs, etc.
1 year ago - mjensenwv said
What will be the implications of this idea?
1
Open Close
How much will we trust fallible human staff office [...]
How much will we trust fallible human staff officers with personal and professional biases to quality-check the work produced by a trusted algorithm? What kind of quality control would they be performing? The benefit of an algorithm is that it could crunch huge amounts of data to produce an outcome far faster than a human could. In order to do any meaningful QC on the output, the human would need to crunch that same data to see the rationale behind decisions being made.

Advanced to the Marketplace.

View this idea
0

When will this idea make a difference?

Your vote Average vote Total votes
Now
2030
2040
2050
2060+

Inspired by

Delegating to Avatar
View Spark
© SciTech Futures 2017. All Rights Reserved.

Site Login