All AI is made up of a large number of algorithms as the essence on which it bases its decision-making. These algorithms are in charge of providing outputs that, when grouped together, determine the presentation of unknown problems or the decision-making in front of a specific situation. Training and the system of sensitivity to change are a fundamental pillar that will determine how these algorithms will interpret and drive their decision-making based on their experience.
All AIs have, as its essence, a large number of algorithms that mark its operation and effectiveness
When AI is being programmed at its earliest stage, it is necessary to establish learning and decision-making mechanisms. These mechanisms are nothing more than interconnected computer algorithms in charge of storing experience to improve decision-making. We can determine that it is important to carry out a correct development of the algorithms for reading and interpreting the data provided, however it is also very important to pay attention to the decision-making algorithms, without ever losing sight of those responsible for making a feed-back to the system for future optimization.
When we program an AI, we have to allow it to establish mechanisms of reaction to change. What seems like a basic concept involves great complexity in its programming since we must choose a more or less sensitive reaction systems. An individual in a peaceful and stable environment should not react too quickly to a change as it is expected to be somewhat punctual and therefore will generate a low error rate with his stable behavior, however an individual who changes or react too quickly to the change will not be so effective. In a stable environment, an individual with a rapid reaction to change would have erratic behavior that would generate a medium-high error rate. On the other hand, if we are faced with an unstable environment, with continuous fluctuations with a certain duration and unpredictability, a quick adaptation capacity will be required that will promptly change its way of making decisions in short periods, waiting for a new adaptation to the new change that nothing, or little, will have to do with any of the previous.
High sensitivity to change can be good or bad in an AI
The effectiveness of an individual, or an AI, will depend on their ability to react to change, high sensitivity is not always good. All AIs require training that is more complicated when the environment is highly unstable and lacks lasting patterns. Establishing a schedule according to each environment, or that establishes a forecast of a ratio of sensitivity to fluctuating change with its consequent impact on how decisions are made, becomes essential.