Most people don’t know that AI isn’t a new term. In fact it was coined by Alan Turing a British logician back in 1930’s. A lot has happened since the 30’s and computer technology has evolved into many areas unrealized or predicted. AI in the past few years is one of those things that has taken on a lot of facets but one thing remains the same and that is security controls need to be applied to ensure business flow accomplishes set goals and companies can meet the high demands of technology and keeping up with competitors. Did you know applying to many controls on AI can almost be worse for some things? On the flip side, not enough can be equally as dangerous. Below is an image from Jason Moccia and it covers the many different layers of AI agents and how they interact with data. Your security model for AI should be well thought out and part of the acceptance criteria for AI implementations. Don’t until later to ask these questions.