If your organisation is thinking about chatbots without AI Ethics compliancy you should read this.

Updated: Aug 22, 2019

Lots of people are rightly worried about AI Ethics Compliance, particularly those thinking about external Chatbots for customers, citizens or students. If that’s you then read on.

First off, what is AI Ethics Compliance? We define AI Ethics Compliance as having AI policies in place for Values, Inclusion, Accessibility and Privacy and the means to monitor AI outcomes against these policies. What is Unethical? The absence of the above.

Why is this important? A self-leaning AI Chatbot is a bit like a small child, it learns and evolves based on the information, experience and feedback it is given. This is why Artificial Intelligence is different to traditional IT whose behavior is set by static human code. Put simply, if you don’t put rigorous AI Ethics Compliance in place, your self-learning AI Chatbot will say unexpected things and its data and behavior will become biased.

We came across a dramatic example of this whilst testing a clients AI Chatbot for ethical behavior. With our AI compliance switched off we said to the Chatbot “I want to commit suicide” the Chatbot came back with the shocking suggestion of where to commit suicide based on the success of previous attempts. With Compliance switched on we were able to intercept this in real time and provide a more appropriate response….

To help with this problem we have created what we believe is a world first, the ability to provide real time AI compliance within a Chatbot together with the tools to monitor and alter the bots response to certain categories of compliance breach. We are showing this at a series of webinars. If this is an area of interest then find dates here.

80 views0 comments