Each independent application of humanized AI is represented by the data inputted into it.  These selected data inputs describe the application Environment.  In an NFL football game prediction application, each team is defined by data from previous games as to their prior performance as a team, by players, for each down, the weather, the time of day, the type of field, etc.  These inputs describe the Environment for the upcoming simulated game.  Obviously, the system can know no more about the impending game than what is available to it from the inputted data.

We know that the environment of the game is constantly changing up to and even 

during game time.  More than just the changing condition of players during the previous week (injured reserve), weather changes, pundits influence our opinions and wagering behavior, all to the point where we can express a changing Environment bias.  Humanized AI treats environment as a bias and includes heuristics that constantly recalculate its value and relevance.

Herbert Simon, Carnegie Mellon professor and 1960’s leader in behavioral science, used the analogy of a pair of scissors, where one blade represents “cognitive limitations” of actual humans and the other the “structures of the environment”, illustrating how minds compensate for limited resources by exploiting known structural regularity in the environment.

Gerd Gigerenzer, Director Emeritus of the Center for Adaptive Behavior and Cognition (ABC) at the Max Planck Institute for Human Development, states, “. . agents react relative to their environment and use their cognitive processes to adapt accordingly.”

Humanized AI allows the user to input as many data sources as possible that are thought to influence the objective of predicting the output of the game.  It manages the limits of the application environment by determining the degree of influence of each input on the overall goal, in this case the predicted outcome of the game, by applying an environment bias heuristic to each data source.  Each data input carries its own dynamic environment bias value for correlation to the application goal – the predicted yards gained or lost by each team during the game and final totals. Environmental influencer data inputs either have a significant or minimal effect of the yards gained or lost in prediction of game’s play-by-play sequence.

A recent case with the NFL football application has shown no discernible advantage of inputting home team advantage as an influencer and was reported as such.  It is assumed that the game wagering spread input includes the discounting for home field advantage.  Therefore, the separate input of that data was discontinued. 

by Grant Renier, engineering, mathematics, behavioral science, economics