Intelligent Agents

Overview
An intelligent agent is system that utilizes information to accomplish some meaningful task. The term intelligent agent can be broken down into multiple parts. An agent is something that is capable of interacting or acting with data or other systems. An intelligent agent is an agent capable of not only interacting with another source but also able to use past experiences to conclude how to best interact with the agents input source. An intelligent agent is heavily reliant on machine learning in order to function as attended.

Generic Agent
A generic agent is an agent that has very limited actions based on a concrete set of inputs. This type of agent has no intelligence and is simply programmed to always react in the same set way to certain conditions. A common example of a generic agent is a thermostat. A thermostat will always receive a temperature as the input and react accordingly. If the room temperature is too cold the thermostat tells the heating device to output more heat, if the room temperature is too hot then the thermostat tells the heating device to lower the output of heat.

Autonomous Agent
An autonomous agent is an agent that is intelligent. Not only does it make decision autonomously based on past experiences but the agent is also able to take into account how it affects its environment and change its decisions based on predictions of the affects its actions will have on its environment.

Reflex Agent
A reflex agent is similar to a generic agent where in the agent only reacts to its environment in the ways it is programmed to. The difference is where as generic agents react the same way every time, a reflexive agent looks up a possible action through a list of rules created for each environmental situation.

Privacy Concerns
Due to the fact that intelligent agents share information there is the concern over user privacy. When an intelligent agent acts upon information received by a user, the data must be transferred and shared between all agents in the process being completed. This means that user data is being shared amongst multiple agents, agent groups and any data sources that the agents may work with. This not only can lead to unauthorized entities getting access to user data but also the possibility that user anonymity could be compromised.