Great article James Duez, thanks for sharing!

You talk about how being able to explain how a judgement is reached is great for the organisations using Rainbird, and their regulators. Do you think there's another potential positive here - the public?

There's a definite worry about how important decisions in individual's life are made by computers. Many people just don't like the idea, whether it's self-driving cars, assessing benefit claims or analysing medical data.

So do you think that embracing explanability in automated decision making also has the potential to make it less 'scary', and maybe improve public acceptance of algorithm-driven choices?

Andrew Cassely This is absolutely better for all customers who lives are impacted by automation every day. It is no longer acceptable that "the computer says no" or what more recently has been called "the AI shrug 🤷‍♂️". Traditional machine learnt approaches to algorithm development lack interpretability because such algorithms are statistical in nature. If you could ask a machine learnt algorithm "why" it made a decision, and if it could answer you, it would say, "you wouldn't understand - you don't speak my language".

We are able to automate complex decisions and describe the rationale for each and every case which gives the author of the system complete confidence that the algorithms are accurate, but also the public the transparency and assurance they need to trust such decisions. This is particularly important when Rainbird is powering digital robo advisors.

Explainability is something Rainbird provides with ease, something both humans and machine learning cannot do - and that is to the benefit of us all!