Qué futuro tiene el futuro? El País reports on our AI research

"We live in an unpredictable time, the leading experts in artifical intelligence tell us. They have no answers and ordinary citizens are not even capable of asking the pertinent questions. We traveled to the Stockholm Institute for Future Studies in search of some light."

This is the preamble to a reportage spanning several pages in the popular weekend supplement to El País, the biggest morning newspaper in Spain. The voices interviewed in this article are five Institute for Futures Studies researchers who are currently taking an interest in artificial intelligence; its variants, how they are spreading, and their impact on democracy, discrimination, labor markets and existence. 

Gustaf Arrhenius, director and professor in practical philosophy, is interviewed. He refers to two timelines we need to keep in mind when discussing the risks of AI. This is partly about the areas wherein AI is already put to use and may cause problems for personal integrity, the creation of false information, expanding surveillance in society or cause inaccuracies in government agency decision making. A markedly longer timeline describes super intelligent AI and its possible future influence on humanity's existence. 

– Personally, more than what the machines are going to do by themselves, I am concerned about what people might be doing with them, Arrhenius says. 

The philosopher Karim Jebari agrees that it is good judgement to differentiate between the two types of AI – the weaker, more specialized variant already in use, and the stronger, more general AI that does not yet exist.

–The problem they can cause are different, but both are worth exploring, says Jebari.

Now that the use of AI has already been popularized, it is not odd that the discussion regarding regulation is arising and gaining traction. This is a point in time where it may be absolutely necessary.

– By nature, regulators are slow and technology moves faster and faster, says the economist and norm specialist Pontus Strimling. Strimling believes it would be a good idea to pause the development of so-called LLM, large language models, such as ChatGPT. – I think it’s a great idea, but not because I’m worried about the extinction of the species, but because it would help us regain a sense of control. The general public, including the political class, feels that technology is something that happens to us, not something we make happen.

The first type of AI is already being used in the field of employee recruitment. Moa Bursell, a sociologist, has been observing the way this enforces the tendencies of discrimination. Her ongoing project has been able to ascertain that, in one company's recruitment process, the use of AI increased discrimination.

– But it wasn’t the algorithm’s fault! Moa Bursell says. The selection the AI made was well-balanced, but when the managers thereafter used it to finalize hiring choices, the result was greater bias than if the managers were involved throughout the process and never relied on any algorithm. – The problem was the human-machine interaction, Bursell concludes.

Since an AI cannot motivate its decisions, the use of it in decision making has risen as a potential democratic problem. The machine becomes a "black box" and is therefore difficult to question. So, is it acceptable to let AI make decisions simply because they are effective at it?

– Democracy demands that decisions be justified before the people. People have the right to know why they have been denied a permit or granted a subsidy, says Ludvig Beckman, professor of political science. – The problem is not that the machines make bad decisions, but rather that these systems that already learn by themselves do so with mechanisms that are not transparent even to the people who program them. 

AI holds the potential to bring forth radical long-term changes for our lives. Will we even be given the opportunity to reject its advances? 

– Where is the user's freedom of choice? wonders Pontus Strimling. And, furthermore, where is the representation of cultural diversity when a very small group of people from a very specific background – mostly developers from Silicon Valley – make decisions without consulting us about what we will all use on a daily basis tomorrow?

Research in which Strimling has studied precisely who these developers are will be published in the near future. 

The article in its entirety is available in both print and behind pay-wall on El País website under the title: "Preguntas muy humanas sobre la inteligencia artificial", authored by Patricia Gosálvez.