Motto: Learning from the past to guide the future
Episode 1
1977 – The GLIA – The first working group on artificial intelligence in Italy
Those who go to sea need a pilot book and in every field of activity to plan and decide you need to have information and points of reference.
My contribution is that of a witness to the history of a term that has conditioned his life and, therefore, the testimony itself. In other words: I am not an objective witness; I am not able to tell the story of artificial intelligence in the last fifty years comprehensively and with due detachment. Instead, I am a biased witness; of the part of those who have seen at times the allied artificial intelligence and, at other times, the enemy.
I was able to thoroughly evaluate them in my discipline sources, the inspirations, the materials that serve to build a solid skill in the field. I learned and helped define the boundaries of what is AI from what is not, the hype, which leads nowhere. In addition to having the “art of computer programming”, to make intelligent systems, experience in the economic sector is needed to evaluate the commitments compared to useless adventures on toys. With this experience it will be possible to understand why, precisely through ignorance of the figures at stake, there are no AI unicorns (a company with a turnover of around one billion) in Italy; how to budget intangibles assets that is to understand why the book value is different from the market value of a company … and also I have learned to reflect on the social consequences of the international redistribution of work resulting from the AI revolution: the formation of new ruling classes, new leadership and. .. new subordinate classes. And again the impact on social media on cyber security …
But let’s see how my adventure started, many years ago, which is also the adventure of Artificial Intelligence in Italy. I was a young programmer of a leading software company, of the IRI group – state holdings – which no longer exists today. We had the State General Accounting Office as an important client, and we took care of the information system. It was the mid-70s ……. The RGS information system had to provide the managers of the Accounting Department with solid information to decide and know how to make forecasts, especially in the State Revenue sector.
The company set up a working group of first-class economists (Romagnoli, Visco, Zaghini, Ferrantini …) for the study of an econometric model capable of forecasting the revenues of the Italian State, divided by income class, by families tax … so as to identify with a certain confidence evasion, avoidance … A very accurate model was developed based on more than 200 differential equations … to solve which it was necessary to invent new theorems and related algorithms, given the computing power and memory available then. This was achieved thanks to a team of first-rate mathematicians: Zirilli, Aluffi Pentini, Parisi, Raganelli to name a few. This very refined and precise model, after the appropriate calibrations, however, had the problem of having to be used by personnel expert in economics but not in computers or software, that is, it had to be able to receive input instructions, structural modification, execution order, of simplified reading of the results and had to know how to respond to the many whys that would come out of the results. The solution implemented involved a conversational module in natural Italian language – edited by the priceless Bruno Graziadio – to communicate directly with the model in all its dimensions of input, structure, results and also answered the many questions about the reason for certain results, perhaps not expected. In fact it was an Italian natural language module that dialogued with a problem solving, this because in the face of instructions in Italian on how and what you wanted to do with the model, the problem solver it produced the pile of machine instructions suitable to solve the request. The expert system set up was called IDIOM = Interactive Dialogue Oriented Machine. It worked so well that a modified version was made for the needs of the newborn SOGEI and its tax inspectors. And it worked so well that the leading economist of the American presidential campaign McGovern, and that Galbraith called one of the most innovative economists of his generation, the econometrician Edwin Kuh, wanted it in English for the Business Forecast courses, at his MIT school, CCREMS, then an integral part of the Sloan School of Management (Wondering what happened to all this in Italy? To get an answer as to why it was used only for internal exercises, you should ask the politicians and bureaucrats of the competent ministries of the time. The technical tool worked very well, but whether to use it or not was the prerogative of the client and political choice). For the development of the natural language module and that of problem solving we had encountered many difficulties, (and maybe we will talk about it again soon, if interested).
It came natural to consult with all those who did AI in Italy, to try to find some idea of a solution. So I began to propose the idea of creating in AICA (Italian Association of Automatic Calculation), then led by Eng. Priests, an interest group on the subject of AI where to exchange information and experiences. This group was born and I have been its coordinator for more than a decade. This group gave me the chance to meet really interesting people.
[[ge:rep-locali:content-hub:306141275]]
AI in Italy started at the beginning of the 70s and in 1977, as I said, the first group of national interest on the sector was born within the AICA (Italian Association of Automatic Calculation): the working group of artificial intelligence (GLIA).
The late Marco Somalvico (General AI and Robotics) and his fantastic group from the Poli of Milan with Guida, Colombetti, Cassinis, the Gini sisters responded to the call with enthusiasm; Gianni Degli Antoni, then at the University of Milan Institute of Cybernetics (Natural language); the Poli of Turin (voice recognition and generation) Pietro Torasso, Renato De Mori and Franco Manucci of Cselt, The great Vincenzo Tagliasco, Adorni, Di Manzo joined immediately from Genoa, who then dealt with general AI, cybernetics, didactics, IR. From Rome the researchers of the cognitive psychology group in via dei Monti Tiburtini, Castelfranchi, Parisi, Graziadio, Lariccia answered; from Pisa, where Father Brusa was active, the very strong Giacomo Ferrari for computational linguistics; from Florence with the Iroe of the CNR Vito Cappellini for electronic imaging and vision and finally Eduardo Caianiello, prestigious cybernetician of the University of Salerno.
The specialties were all there: the party could begin.
Italy was at that point aligned with the other industrialized countries in accepting the scientific and industrial challenge of AI. But immediately various elements of structural weakness began to delay and progressively limit the development of the sector, both from the point of view of research and in subsequent years and in a more marked way, from that of technology transfer and applications.
About thirty researchers in the group, then from 1977, which later became 150 in 1981 and 955 in 1987.
End of the 1st episode
In the next episodes:
2- The Roaring Twenties or when the centers that matter talked to companies and not to individual researchers: Olivetti in Cupertino, Tecsiel at MIT, Carnegie Mellon, Yale, Stanford Research Institute
3- Milan, the world capital of AI : IJCAI 1987,
4-Seen up close: gli incontri con Bill Gates, Steve Jobs, Marvin Minsky, Nicholas Negroponte, Carl Hewitt, Gordon Plotkin …
.