Natural language understanding is in particular an activity required by about 76% of the work activities analysed by McKinsey. Improving natural language understanding capabilities would have an impact on the equivalent of $3 trillion in wages.
The biggest impact on expanding the number of work activities that machine learning could technically automatize include the three capabilities: natural language understanding (76%), recognizing known patterns (99%), natural language generation (79%).
Routine Tasks Will Be Automated
The occupations that involve the highest level of time spent on activities that could be automated if machine learning improves include customer service representatives, first-line supervisors of office and administrative support workers, lawyers, business operation specialists, as well as secretary and administrative assistants.
Does this mean that the above jobs will disappear? No, the implication is that a large part of the activities performed by these jobs will be automated, leading to a focus on tasks that truly require human skills, creativity and reasoning. Only the routine tasks will be automated, but bringing massive productivity gains.
Besides the technical challenges in bringing the technology to a state where automation is indeed possible, there are other inherent challenges including the fact that machine learning models are in many cases opaque and decision making on the basis of machine learned models might be intransparent as well as the fact that machine learning algorithms could acquire unwanted or even discriminatory biases.
http://www.semalytix.de/