“We need to reflect on what we have created and why.” They are the words of Nello Cristianini, physicist, professor and researcher of Artificial Intelligence at the University of Bristol, during the thirty-sixth edition of the Reading of the Mill entitled “Intelligent machines. Understand them to be able to live with them”.
From the 50s to the 90s, the approach to artificial intelligence was very "mathematical": one wondered what was the highest intellectual activity that a human could perform and the answer seemed to be "prove theorems", he began the Bristol lecturer." The idea that intelligence was logical deduction, and axiomatic rules was very popular in those days, and all the efforts invested in trying to create intelligent machines in this vein had led to no results. The turning point was the abandonment of this approach in favor of statistical methods: this "paradigm shift" allowed for the first translators. In summary, we have gone from trying to define precise rules and proceeding by logical deduction to following a behavioral model guided by statistical inference.
We have avoided taking the human being as a model of intelligence: the idea is to be interested in intelligent behavior rather than in intelligence itself, which can also be typical of a system. The concept is to emulate useful behaviors, learn from examples, leaving aside the attempt to solve the "intelligence problem", also abandoning the question of giving a label to the concept itself or defining it unequivocally. Amazon manages to advise without understanding the model of a personality, in the same way it is possible to block unwanted emails.
However, in addition to all the advantages that these machines have brought, there are real and concrete risks not only to our existence and freedom, but also to privacy. “Public opinion itself is conditioned by algorithms” underlined Cristianini. “We have created machines that we can no longer do without and we still don't have the cultural norms, laws and values to live with”. It is a very topical problem of cultural structure. There is a need for people who "paint" the new problems to pass on to future generations an awareness of these risks, and this task, for Nello, belongs to politics.
It is a question of governing change while being aware of the many implications: in the educational field, at a social level but also on our mind. Machines are a powerful and essential tool, just think of the many achievements achieved in the medical, industrial and scientific fields. However, we need to think about the ethical and cultural consequences that the large-scale use of intelligent machines will have. This implies the need for legislative changes to clarify roles and responsibilities.
The problem is not the machine, but man himself if he doesn't address issues from the right perspective. One wonders if in this process of replacing man what are the ethical and cultural consequences on a global level. And the risk for man of losing part of his identity, reducing his own existence to what a machine can do and to the satisfaction of new, mostly induced needs.
The European Parliament has expressed itself on the matter: "The use ofartificial intelligence (AI) it must prevent gender, social and cultural prejudices and protect linguistic diversity, cornerstones of the fundamental rights and values of our society". With 623 votes in favor from members who want to underline the importance of European ethical values at the basis of the use, in educational and cultural fields, of technologies developed thanks to AI.
With this resolution, the European Parliament "invites the European Commission and all Member States to take into account ethical aspects, including from a gender perspective, in the development of policies and legislation on AI and, if necessary, to adapt the current legislation, including Union programs and ethical guidelines on the use of AI".
In particular, the parliamentarians call for the development of specific indicators to measure diversity and ensure that the recommendations contained in the algorithms, especially in audio and video streaming services, do not negatively affect the cultural and linguistic diversity of the Union.
Attention is also placed on the diffusion of fake news, a hot topic that Professor Cristianini also underlined during the event. We need to establish how artificial intelligence technologies are used by the media and be vigilant about the misuse of these technologies to spread false news.
