Cybercrime – Artificial Intelligence.

Statement: An ‘Artificial Intelligence’ or machine learning system should be prohibited from making a serious (non-trivial) decision about a person UNLESS it is able to generate a comprehensible explanation of the reasons for how it reached that decision, including the rules and facts it found relevant.

Yes or No to the statement and Why?

If Yes,

These details need to be included:

  1. The facts and evidence which were taken into account
  2. The ‘legal’ or administrative reasons for the decision, in terms of:
  • the rules, criteria, principles etc. which were considered to apply in that situation, and
  • how the rules were applied to the facts to reach a series of conclusions, and thus the ultimate decision.