Machine ethicists are looking for psychologists, philosophers, and linguists to "educate" robots

This article was written and published in Spanish and has been translated into English via Google Translate. Click here to read the original article.

Is it possible to teach ethics to a machine?

Philosophers, psychologists, sociologists and linguists are part of multidisciplinary teams that work to avoid possible malicious, immoral or discriminatory uses of artificial intelligence.

The end of 2019 is approaching and next to it come infinite compilations with the highlights of the year. From a technological perspective, the last twelve months have a clear protagonist: artificial intelligence (AI). Richard Benjamins, data and AI ambassador for LUCA, Telefónica's data unit, highlights three recent milestones that will mark a before and after in the evolution of this technology.

Why is 2019 the year of artificial intelligence?

AI has existed since the 50s, but now is when it is really adding value thanks to the availability of large amounts of data and the increase in the processing capacity of the machines, to which more and more automated tasks are entrusted. Despite this unstoppable progress, Benjamins emphasizes that when talking about AI it is necessary to differentiate between science (what is possible), fiction (what is still far from being achieved) and opinion (the personal judgment of each individual).

"People have emotions and we can reflect, but a machine all it does is execute things and it doesn't really know what decisions it is making."

Do machines make better decisions than people?

One of the main technological challenges of today is to prevent AI from repeating the injustices of society.

"It's about managing your risks and taking advantage of your opportunities, for which you have to eliminate any possible bias in the data used to train the algorithms."

In this sense, Telefónica Empresas has published its Manifesto for a new digital pact, where it expresses that people-centered technology must ensure that its main beneficiaries are citizens.

How is ethics taught to a machine?

To make the algorithms as impartial as possible, a methodology is needed to help apply certain AI principles. Telefónica Empresas has designed a model that can be followed by organizations involved in the development of projects that include artificial intelligence. According to Benjamins, this methodology consists of five elements: the principles of AI itself, a continuous training of all professionals, self-assessment questionnaires that must be completed by those involved in an AI project to ensure that it has been defined according to the principles, technical tools that help detect possible biases, and a governance model that assigns responsibilities when solving incidents.

Who controls those who launch artificial intelligence projects?

In short, the objective is that projects combining artificial intelligence with other technologies are sustainable from a commercial, social, and ethical perspective, for which it is necessary to have templates prepared for this challenge.

This article was written and published in Spanish and has been translated into English via Google Translate. Click here to read the original article.

Join us February 26-27 for the Property Portal Watch Conference Bangkok 2020.