IMAGE: Leo Blanchette — 123RF

Man and algorithms

A article in MIT Technology Review, “Your best teammate might someday be an algorithm”, prompts thoughts about the balance between machines developing more and more capabilities that will doubtless take away many jobs, and my own vision of my work, which is increasingly improved by the use of such technologies.

Frankly, I’ve never been too afraid that an algorithm will take away my job. In large part that is because I am fortunate enough to work in an area that for the moment does not appear to be in danger of being taken over by robots, along with the total flexibility that the institution I work for gives me, but I have always thought that if education were revolutionized by technology, I would always be able to study it and explain it to someone who could add value to it.

Thus, algorithmic tool development focused on collaboration makes me wonder what things I would be able to do when some of the parts of my work I least like and that add less value can be carried out by algorithms that will be increasingly easier to develop and educate thanks to simpler tools.

We are increasingly surrounded by tools that make work and our personal lives easier. That said, we are still a long way from the full automation of many tasks and that require what we might call a human touch, meaning they fail to perform at the required level of performance.

For many years, automation has been a way to replicate processes in an identical and predictable way, an approach that has its limitations. Introducing into these algorithms steps that replicate at certain points human judgment based on previously generated data would allow us to create much more powerful automation based on a range of parameters that would add value. Until recently, separating spam from email was a tiresome manual task. Now, practically nothing escapes the algorithm and there is virtually no need to check our spam folder. The same applies to a thousand other things.

Many of the tasks I carry out throughout the day require human attributes. But many are more mechanical and repetitive and would clearly benefit from algorithmic attention. Before robots steal our jobs, we will see many other cases of algorithmic development that will increase automation way beyond today’s levels, and applied to tasks that while not completely repetitive, certainly seem so, and that will enhance our capabilities while freeing up time for other tasks. In many cases, by the time algorithms and robots are able to do our work, we will already be doing different jobs, following previous examples in history, and that seem to me more constructive and logical. The more algorithms I see, the more I want to work with them, to apply them to my daily activities, and enhance my abilities as a human being.

(En español, aquí)