HackingTheUniverse

The day is coming soon when we will use software agents that do integrity checking just as readily as we now use spell checkers. For an individual or group to be considered in a state of high integrity, we require behavior (actions taken) that is consistent with a framework of principles or ethic values. In order to create a functional integrity checker, we need two basic sections:

1. An acceptable definition of the framework, subdivided as follows:
   • Facts
     • Simple facts that are easy to check will require little or no translation
     • Complex and compound facts that are not easy to check will require a massaging process to get them into formats that can be easily checked
   • Limits, boundaries and other requirements
     • Maximums, minimums and other simple threshholds will be easy
     • “Shapes and flavors and tastes” – meaning a variety of conditions that are more complex, come in many different forms and will require translation before comparison
   • Principles, morals and ethical conditions
2. An engine that can check a body of information against that framework:
   • Simple comparisons – number or color or shape equivalency determination
   • Multi-valued logic comparisons – will require more sophisticated mathematics, neural network technology, analog circuitry…?
   • Self-learning and improving capability – some of this will be as simple as prompting the user to create a new “field” or “column” in the database and populate it with input requirements or fill-in help. Neural network self-learning capability can offer wise suggestions for the user to accept or reject.

Just as it is possible to use different dictionary files with a spell-checker, it will be necessary to use different “framework” base configurations with different knowledge domains.

SEE ALSO:
The Logic of Ethics
Ethical Advisors
Ethics as Prediction
AI Trust Issues