Automation brings complexity, but automation can also bring simplifications in a process. This because with automation we need to create a model, write assumptions and exceptions and define of course requirements which are feasibly for implementation.
Nowadays a dominant thinking trend is visible that everything can be easily automated. By computers, robots and with use of AI software. However the question of what and how to automate is within a specific context not always that simple.
Long long time ago simple list of MABA-MABA appeared. MABA-MABA stands for “Men Are Better At -vs- Machines Are Better At”. In essence computers and robots are good at simple repetitive task where no reasoning on context related information is needed to produce a good outcome. But now with all new emerging AI algorithms based on Big Data analyzes are appearing some assumptions on where automation is good and where not need to be revisited. Car driving, medical diagnoses, legal advice are all fields where a shift is being made towards more automation.
Also in the automation of IT more and more once complex IT work is automated. We see software defined data centers, software controlled automated delivery pipelines and more and better tools that prevent accidents within the automation creation process to occur. Does this mean that there is no place left for IT engineers or architects with deep technical knowledge? I think humans are always needed to control and audit the (automated) QA processes. Whether on security, privacy or old school software reliability and safety. However since the majority of IT projects is still slow in picking up all the current possibilities available for automation of the IT work flow. So a fast shift will be hard due to all the factors that limits humans to adapt a radically fast change a current way of working. But the emphasize on skills that really bring extra value for a company will change over time.