The key to "conscious unbias" in AI algorithms: Graph Analytics

AI is not only perpetuating our unconscious biases - it is also concealing them. When AI algorithms are so hard to unpick, we are even less likely to uncover the inbuilt prejudices that were never intended but that continue to have a detrimental impact. The key to an ethically sound intelligent operation is an intentional and transparent AI whose moral conscience is able to evolve in parallel with our own - and we have an opportunity to do this by ensuring every AI references the appropriate social context for its calculations and decisioning.
Who is this presentation for?
Business Leaders, Team Leaders, Senior Management, Data and Analytics Leaders, HR teams and anyone running AI or Machine Learning models where data might contain unconscious bias
Prerequisite knowledge:
Basic consumer knowledge of Machine Learning and, Imagine Recognition and Natural Language Processing
What you'll learn?
1. Unconscious bias permeates even the most seemingly neutral algorithms - and this creates real future organisational risk 2. If set up appropriately, we can consciously interrogate and iterate algorithms that require additional ethical considerations or re-wiring 3. By applying knowledge graphs to AI algorithms, we can ensure the appropriate social context is factored into calculations to automatically mitigate bias

Charlie Beveridge is an Independent Advisor and SME in Graph Analytics & Digital Transformation, currently on sabbatical from her position as a Senior Manager at Accenture Strategy and Consulting. At Accenture, Charlie led their AI propositions for clients in the Telco and Media industries. Charlie specialises in the implementation of disruptive AI technologies that transform capabilities in operations, HR, service, sales and marketing teams. Prior to joining Accenture, Charlie worked in IT Advisory at PA Consulting.