EY Asks: How Inclusive Is Your AI Solution?
Challenge
Despite dramatic advancements in Artificial Intelligence, studies confirm that facial recognition systems are even more unreliable and racially biased than we feared. This is a result of using datasets that don’t represent all of humanity to train machine learning models. Implementing AI without first asking the right questions could lead to negative consequences for consumers, businesses and the public sector.
Idea
We want to provoke conversation on the topic of inclusive technology by giving people a digital out-of-home experience of first being excluded and then included by tech – “the medium is the message”. This positions EY as the AI consultancy bold enough to ask difficult questions.
Words on the digital screen/billboard
Headline: How do we ensure AI doesn’t leave anyone behind?
Endline: Reframe your future. EY.
Making people invisible
Using some amazing TF.js code, I was able to turn one or more people invisible in a video live stream. If you look carefully you’ll see that non-human objects will still be seen moving in the background as normal - adding to the magic of the experience. A few experiential ideas using this clever technology was brainstormed and one was even presented to a client.