RSS feed Add to your Facebook page Watch us on Youtube

Activity title

Digital Employees for Network Management and Control

Activity Reference

IST-ET-105 (AI2S)

Panel

Information Systems Technology

Security Classification

PUBLIC RELEASE

Status

Planning

Activity type

ET

Start date

2018

End date

2019

Keywords

AI, Chat Bots, Cognitive Agents, Machine Learning, Network Management and Control

Background

In the commercial market there have recently emerged a number of software tools that are aimed at ‘off-loading’ many of the tasks performed by the traditional IT Service Desk. These Intelligent Agents (IAs) are purported to be capable of answering the customers’ calls (or chats, etc.) and taking actions to resolve issues. These IAs are apparently able to handle many simultaneous sessions, can learn from their experience, are able to digest documentation and manuals defining the networks, and take action as authorised to deal with the issue or escalate it to an appropriate human [resolver cell] automatically. These agents can be contextually aware, that is, can build up a knowledge of the underlying network configuration, including its current state, and can thus alter their decision logic based on this knowledge. The potential exists for these agents to be used in the military to support both the static and deployed networks. They might lower costs, reduce call resolution times, and increase customer satisfaction. They can be employed 24/7, do not get tired or lose effectiveness after many months in stressful situations and can scale to handle any number of simultaneous calls. They thus offer great potential in releasing valuable [military or civilian] manpower that can be repurposed for higher level tasks, or to simply save money. However, they may also present some disadvantages. In order to be effective, they need to be trusted and be given some level of administrative privileges. As they are able to learn, they might present a new cyber-attack vector by being taught behaviours disadvantageous to our military. This may present problems with accreditation and trust. Nevertheless, this is likely to be an area where we first see the wide-scale adoption of Artificial Intelligence in our military domain as the products available from the commercial IT market are likely to align well with the needs of the military. If adopted these agents also represent what will likely be the military’s first off-loading of human and civilian defence employees to ‘digital employees’ which will potentially raise ethical, moral and legal issues.

Objectives

This activity will assess the potential of these technologies for use on defence networks, assess their potential advantages and disadvantages. It will produce and initial feasibility study into deploying digital employees to carry out SMC functions in deployed and static locations.

Topics

To begin, the ET will conduct a market survey, to understand the offerings available on the market. A comparative assessment will be made that attempts to categorise the various products based on their feature sets to give a broad understanding of the capabilities and maturity of the products. This may involve interaction / demonstrations with and by the vendors. Based on what is discovered in the market survey, a ‘skeletal’ business case will be assembled, that will determine at a high level the benefits and dis-benefits of the technology, the risks, the likely timeframes for adoption, etc. It will address financial aspects – cost versus benefit – and also the cyber security aspects. Although the study will be focussed on the service management and control aspects of this technology, the study will also document any related non-technical aspects related to this type of technology that are discovered or arise; that is, for example, the displacement of a segment of the human workforce with intelligent agents, in particular in the military domain.

Contact Panel Office