Automated testing is a liability. Manual checking functionality is a huge liability. I want to emphasize some concerns I have with the machine learning testing solutions.
If you are not familiar with testing vs checking here is my quick explanation. Computers perform checking, humans when they are not following a script perform
The primary issue I take with the machine learning automation is it tends to focus solely on web frontend verification. For certain websites this my be sufficient, however I think that is likely very few. In an earlier article I described a web UI testing setup I'm using.
Machine learning would not allow me to use this method of control. It also does not focus on other validation like database reporting.
There are a lot of moving parts when it comes to a "working" application. If your test strategy could be replaced by the machine learning tools of today, then it I claim that wasn't the right strategy to begin with.
However machine learning is a valuable endeavor and could use some funding. Even though I don't view it as the right strategy, it is better than nothing and perpetuates the most common test strategy.