We are thrilled by the number of participants joining our training on AI Basics for Testers. We have more than 2500 testers from the community, which includes Test Architects, Sr Managers, Data Scientists and CTO’s participating in the training sessions. We had the first session on Saturday 13th May 2017, 8 sessions lined up every week (Saturday), 1 session per week.
With respect to these sessions, we have been receiving many questions from the participants which I will try to answer here.
Why is it important for testers to learn AI?
The fundamental process of developing traditional software which includes Web, Desktop, of Mobile apps are mostly similar and follows common development life cycle. Given below is one of the development model and comparisons with how AI systems are built up
These two methods differ because Machine Learning systems evolve with Data. If the development process is different, why does it need different testing methods?
As Machine Learning systems depend on data and algorithms to work – traditional testing methods that focus on functionality and coverage need to adapt to changes as to how these products are built.
Google has just announced Google Home in Sundar Pichai’s keynote, Mark Zukerberg announced Jarvis and how it was built, Tesla has been successful in building self-driving cars, giving us a hint about how Machine Learning systems are going to change the world. However, there is still a challenge in testing these systems for which testers and organizations need to upskill themselves so that they can test these systems better.
Why is it important for the world that testers learn AI?
Right now there are multiple reasons why testers cannot ignore learning about AI. Machine learning applications are not 100% accurate – and almost never will be. The most fundamental reason is that this application learning is limited by data they have used to build algorithms. Machine learning apps are managing almost every daily activity that is performed by humans – one error can lead to serious damages. False Positives are a big problem ( Medical field is using this term from decades ) – and products which deal with human life – cars, medical instruments, etc, need to be correct at times. This problem has been illustrated by Peter Thiel in his book Zero to One. ( A really good read ) .
I will take a little break now – as I need to prepare for the next session in Data Science and Machine Learning. If you have missed it – please watch it on our official YouTube channel.
Author – Riyaj Shaikh | Chief Data Officer | Moolya1