Me, myself & AI: in the OECD forum
Yesterday I had the opportunity to participate as a discussion leader in a session entitled “Me, myself & AI” within the Forum of the Organization for Economic Cooperation and Development (OECD): three hours of very interesting discussion in an intensely participatory format, during which we discussed the evolution and the future of a world where machine learning algorithms are playing a growing role.
Spanish news agency Europa Press reported some of my comments (link in Spanish) during the forum, covered today in the Spanish media: the need for businesses to shift their activities toward obtaining data to feed algorithms that provide efficient, fast and differential management capacity. We are still a long way from the development of artificial general intelligence as such, but machine learning algorithms are increasingly positioning themselves as optimal solutions for a growing number of tasks, and are increasingly accessible. The emergence of machine learning as a Service (MLaaS) platforms and the decision by big technology companies to position themselves in this area through the acquisition of startups active in the field are creating a scenario that is increasingly influential when it comes to establishing companies’ competitiveness: machine learning algorithms define what products are offered to which potential customers, pricing policies, churn prevention, detection of possible fraudulent activity, options in the determination of the marketing mix, and many more things.
But beyond these issues, which are exercising growing influence on companies, and without peddling the myth that we are already surrounded by artificial intelligence and, soon, for sure, killer robots, the opportunity to participate in a forum like yesterday’s, with specialists from different companies, people who have been developing language processing or computer vision algorithms for years, or with politicians who are making decisions that condition companies’ scope is very interesting.
Among the issues we discussed at the forum were, among others, how to structure the dependent variables an algorithm must optimize (and what it means to optimize them), the need to think about managing a new operating system in which many significant aspects of human activity will be developed, or the interaction between algorithms and people.
In my presentation, my approach was to see all this in terms of language programming: what will be the variables to be defined and how to define them, how to structure conditionals and how to generate loops to control the resulting activity, as well as touching on the skills we will have to develop, all in an interview I will publish as soon as it becomes available. There were also interesting discussions about ethical issues, possible conflicts between the decisions of an algorithm and those of a person, or possible biases that may arise.
These OECD forums are of great importance for the development of lines of thought that will influence the legislative environment and the decisions of companies, governments and organizations of all kinds, including the ethics behind them. I am very grateful to the OECD for the opportunity to participate. As soon as I have more materials available, I will add them.
(En español, aquí)
