CASTx17

At Xero we’re using Machine Learning and A.I to make accounting more accessible to small business owners by predicting account codes and providing an always accessible chatbot to answer those tricky questions.

What does this all mean for the human development teams behind the machines?

If the goal of Machine Learning is to make decisions without the need for human intervention, where does that leave us?

Fortunately, this doesn’t mean that testing is no longer required but it does mean that testing is changing. It means that we need to understand the human interactions with our software more now than ever before. By looking at the output of results of an algorithm under test, we can help teach the software how to address unexpected inputs and also understand which algorithm is best suited to our goals.

Just like humans, while machine’s are learning, mistakes are bound to be made. This is why human intervention is so important throughout the learning process. By explaining the different phases of learning concentrating on Supervised Learning leading to Unsupervised Learning, I will show how being a creatively unique human thinker can prevent unfavourable decisions being made by the algorithms.

To conclude, I will be drawing parallels with the Industrial Revolution and the introduction of manufacturing that made production quicker. I will also show how there will always be a need for those of us passionate about our craft.

The introduction of the printing press is estimated to have produced as many books in the following 50 years as were produced in the previous 1200. With the use of machine learning we’re seeing more decisions being made faster than ever before. If decision making is the currency of the A.I, then understanding and being uniquely human is the currency of us.