AGI – Artificial intelligence: how much confusion two words can make. First of all because there is nothing intelligent. And then because the combination is deceptive: it seems the fruit of a marriage, the one between engineering and biology, united in the attempts to recreate the human mind. “It’s actually a divorce,” says Luciano Floridi, professor of Philosophy and Information Ethics at Oxford and president of the IFAB International Scientific Board. No drama: a divorce can be a step towards creating a marriage between “blue and green”. That is, between digital and environmental protection.
Professor, why is artificial intelligence a divorce?
Why creating artifacts and trying to replicate human intelligence are actually divergent. With artificial intelligence divorce is the ability to act successfully and the need to be intelligent to do so. I give an example …
You are welcome…
The robot that cuts the grass, moving in the garden in total autonomy, has zero intelligence, but it does one thing very well. If I were to do it myself, I would be less efficient and would have to apply my intelligence to avoid destroying the rosebush. It is a miracle: it has never happened before to have great success in something without applying intelligence. In fact, today we produce extraordinary washing machines, however capable of doing increasingly complex things.
How do you make this divorce work?
Divorce works when it is possible to “untie” process and intelligence. Success therefore depends on how well I can transform the environment to be welcoming to AI. Let’s look at self-driving cars: they already manage to go from A to B, but to work they have to change the roads. The traffic light is made for our eyes, while artificial intelligence needs a radio transmitter. Thanks to this transformation process, divorce can work: the problem must be rethought so that it can be solved by an artificial intelligence.
Shaping the environment for artificial intelligence opens up ethical issues and potential risks …
Yes sure. Ability to act and intelligence have always traveled together. And in the gap created by this unprecedented disconnect, problems arise. The first concerns the value of people and our autonomy. It is dangerous, for example, that there are men managed by software in the gig economy. It is the beginning of a dystopian society we don’t want to live in. But our autonomy is also at stake when we watch a movie in streaming: if you have watched A, the platform suggests you to see B and C, influencing your choices. Then there is also another problem.
Which?
They are the biases [“pregiudizi” che un algoritmo tende a rafforzare, ndr]. If software makes unfair decisions, don’t blame artificial intelligence but data. Artificial intelligence has no ability to correct for biases. If a company has always hired men, managers can understand that this is bad behavior and decide to hire women as well. But if an artificial intelligence is hired to select candidates based on the data, it will continue to indicate men.
Is that why you say that the real challenge is not digital innovation but governance?
What makes the difference is who has his hands on the wheel, not who has his foot on the accelerator. European legislation, for example, says there must be constant human supervision.
Speaking of European legislation, what do you think of the regulation that should regulate artificial intelligence services, the so-called “AI Act”?
The EU is doing an excellent job. I am optimistic about the direction and have to be patient about the timing. The philosopher can also see things first, but he must know how to wait. For the GDPR, for example, it took six years. It is a delicate course, for a marathon runner: it cannot be compared with 100 meters. When, within a couple of years, we have the AI Act, the EU will have a leadership role.
Does the fact that this is a Community regulation and not a global one risk weakening it?
I believe it will be very effective even if not global. It is called the “Brussels effect”: the EU has influence over other countries because the legislation raises standards and requests for compliance. Large companies are adapting and not only for the EU market, because it is not convenient to have two different protocols. This generates a factual impact even before a legal one. We saw it with the GDPR: the EU detaches itself, the rest of the group adjusts due to inertia. The European one thus becomes a “leadership by example”. I expect EU legislation to have enormous influence within five years, not just in Europe.
The impact of digital goes beyond single technology and regulatory frameworks. In one of your works you speak of the “fourth revolution”.
With Turing, computer science and artificial intelligence, man has been displaced for the fourth time. Copernicus moved us from the center of the universe, Darwin told us that we are not at the center of the animal kingdom and Freud that we are not at the center of our mind. Today we feel the pressure because we are no longer even at the center of the infosphere, the space of data, information and knowledge. We should then ask ourselves “why is humanity special?” and rethink the foundations of philosophical anthropology, based on exceptionalism. We should develop a philosophy where we are not at the center of anything. True human capacity is to be “in the service of”.
For example at the service of the environment: if artificial intelligence is a divorce, in his latest book he proposes a marriage, between “green and blue” …
On the one hand we have AI, an extraordinary, flexible force with a great ability to solve problems; on the other hand we have many social and environmental problems to solve. Let’s put one and the other together, at the service of sustainability and a more equitable society, which knows not only to generate wealth but also to distribute it. With the blue of digital technologies and the green of the environment we can do more with less, differently and better.
What are or will be the closest applications of this marriage?
Already now, thanks to modeling and data analysis, it is possible to better understand the functioning of public transport, the energy consumption of a city, the impact of infrastructures and new buildings. Knowing more allows you to draw much better what you are designing. An example is that of the “twin c” iy, a digital replica of a city that allows you to study traffic to see how it changes based on some variables.
The link between artificial intelligence and the environment is one of the areas on which the International Foundation Big Data and Artificial Intelligence for Human Development (Ifab) aims …
No one, alone, has the keys to these problems. Public and private collaboration, managerial skills, technological resources, policies, data and the consent of citizens are needed. Today these skills are found in different places. Ifab understood the importance of aggregating them, creating the Big Data Technopole hub. It is a difficult challenge, but one that is worth taking up: it is the way to use divorce to make it a marriage.
.