It seemed like an ordinary speech on the agreements between Italy and Switzerland for cross-border workers. Up to the surprise conclusion: «The speech you have just listened to is not mine nor the product of human intelligence».
Marco Lombardo, senator of Azione-Italia viva, was the first parliamentarian in Italy to read a text entirely written by artificial intelligence in the classroom.
The intervention contained information on funds, subjective evaluations, political proposals. How did he create it with AI?
«With a company that operates in this area, I provided the Chat GPT 4 artificial intelligence software with the draft law of the Italian-Swiss agreement on cross-border workers and the interventions that had been made on the subject both in the Chamber and in the Senate by the group Action-Italy alive. Thus the program has created a flow of thought on the basis of the documents that have been sent to it, with its own elaborations. No manipulation has been done on the text produced».
Why this specific topic?
«It seemed useful to show that artificial intelligence algorithms do not apply only to technological processes or digital-related issues, but potentially to any bill».
How did you come up with the idea of doing an intervention generated by artificial intelligence?
«I've been dealing with how algorithms are transforming the world of work for some time. And therefore also the ethical, economic and social implications. It seemed important to me that the Italian parliament also open its eyes to a phenomenon that is now unavoidable".
The populist question would be: so we no longer need parliamentarians?
«I also wanted to make political decision-makers and parliamentary colleagues understand that their work too can be "threatened" by artificial intelligence. Not even politics can think of exempting itself from a comparison with algorithms. You need to know how to use it consciously. If everything is reduced to a question of cost and efficiency, our role could also be replaced. We have to put something more into it: head, heart, ideas».
In the courtroom the text written by artificial intelligence: here is the intervention of the senator from Action
What are the opportunities, what are the risks of this technology?
“The opportunities are enormous. Think of all the applications on precision agriculture, on the possibility of saving natural resources thanks to a more targeted use. Already today in some sectors such as insurance and banking they resort to predictive systems. Then there are the risks: the algorithms are based on human inputs. And therefore they carry out analyzes that can also have discriminatory effects. In addition, there are implications for journalism, as we have seen, no longer having to deal with fake news, but with deepfakes, images that are very difficult to distinguish".
Do you think that the Italian parliament can or should intervene with a law, as indeed is being done at European level?
«If we start with a regulatory approach without knowing the phenomenon, we risk slowing down development and not being effective in preventing risks. Many fellow parliamentarians were intrigued at the end of my speech. There is a need for a lot of information first, at all levels. When we have reached an important degree of maturity, we will be ready to do a series of hearings with public and private entities, as in the EU and the US, which can help contain the risks and exploit the greatest potential".
In Italy, the first move of the Privacy Guarantor - with subsequent rethinking - was to ban Chat GPT. An attitude that risks keeping us out of a crucial game?
«Yes, it also applies to the EU. Europe cannot limit itself to acting as referee because by definition the referee never wins. He must decide to be a player, a player on the pitch, both nationally and supranationally. Because when we talk about sovereignty and democracy, we cannot fail to talk about digital democracy".
In a letter, the fathers of artificial intelligence sounded an apocalyptic alarm: «Mitigating the risk of extinction due to AI should be a global priority». Excessive?
«I think that even the tones of that warning serve to raise the level of attention. However, the issues that pose not only social but also ethical consequences are important. We think of a potential use of artificial intelligence to alter decision-making processes, perhaps at the hands of foreign authorities. We must not demonize artificial intelligence and algorithms, on the contrary. But the machine that must be at the service of man and not be a threat to democracy».
Could the AI also have conducted this interview?
“Probably yes, absolutely. By supplying the software with my speeches from today and from before».
And if so, would the reader be able to notice?
“Very difficult. This is why human intelligence, with its creativity and ability to create original connections, must remain a distinctive trait of any activity, even parliamentary one".