For a Human-Centered AI

MISLEADING MACHINES. ARTIFICIAL INTELLIGENCE AFTER THE TURING TEST

April 22, 2020

The first webinar organized by FBK-ISIG in the time of Covid-19 focused on the analysis by media historian Simone Natale

(v.l.) Interestingly, Simone Natale, a historian at Loughborough University (UK) opens his lecture by discussing the seances of the Victorian era to investigate the field of artificial intelligence studies.

There are more than 60 of us, on Friday, April 17, around the virtual table of the seminar entitled “Misleading machines. Artificial Intelligence after the Turing test”, the first one organized with remote videoconferencing systems, as pointed out in the introduction by director Christoph Cornelissen and by researcher Maurizio Cau, from Fondazione Bruno Kessler’s Italian-German Historical Institute.

“It was British scientist Michael Faraday,” Natale explained, “who, as an expert in electromagnetism, decided to investigate the phenomenon of spiritualism. However, he did not do it only from a specialist point of view. Rather, he worked on the hypothesis that events had to do with the participation of people, who, through deception or unconsciously, caused phenomena, such as table moving”.

And it is precisely the people, their perceptions and their reactions that are at the center of the scholar’s analysis also in the approach with artificial intelligence, especially from the moment it is programmed to communicate with us. So much so that even for the Turing test, one of the founding acts of artificial intelligence, the question of humans influencing the test emerges, through their higher or lower IT skills proficiency or their possible prejudices.

Since its origins in the mid-twentieth century, artificial intelligence has been nourished by the dream – cultivated by some scholars and considered unrealistic by others – to create forms of consciousness similar or equivalent to that of human beings. Yet, according to Natale, it could be described more accurately as a set of technologies that produce the appearance of an intellect. Through the analysis of some aspects of the history of artificial intelligence, from its origins to voice assistants Siri and Alexa, Simone Natale proposes to overturn the discourse on these technologies, suggesting that the “intelligence” of the machines lies first of all in the perception of humans interacting with them. “Adopting this point of view will help us realize to what extent our tendency to project humanity and intelligence onto objects, combined with information technologies, can transform the essence of sociability and everyday life”.

Basically, approaching this field of investigation not only with an engineering approach but – like Faraday did for spiritualism – with a broader view is critical, so as to consider the many socio-cultural implications that we put into action as human beings.

So as not to end up asking Siri if it believes in spirits and maybe being replied: “I don’t think I have beliefs”.


The author/s