Artificial Intelligence (AI) and Ground Rules

Given the impressive and ever more impressive results of AI there are many well publicised efforts being made to draw up standards and even, perhaps, enforce them. The standards all cover the data protection element - the acquisition and processing (in data security jargon) of information - and the computer programs and programming. There is need for audit trails. Much the most important element governs the use and decision making by humans.

AI 'machines' have no agency. Could one turn itself on one morning and say something like "I'm bored. I'm fed up playing chess. Today I'm going to .... instead." Not really, although it might be possible to fake it by including an algorithm that generates a 'random' flag that channels the AI 'machine's activity into different programs, but no more.

Nevertheless, all of this activity on AI gets us thinking about how we humans would like thinking machines to behave and what controls we may want to exert over them.