Artificial intelligence is here to stay and they know it well in the European Union. That seems to be one of the messages coming out of the European Parliament at a time when, moreover, the machinery is already underway to legislate an area that is full of surprises.. He talks about this for 20 minutes with Ibán García del Blanco (León, 1977), a PSOE MEP, who, like many of his colleagues, closely follows the evolution of this new reality and asks, above all, to understand it in order to adapt to it smoothly.
Is artificial intelligence a risk or an opportunity? I think that it is fundamentally an opportunity that, due to its size, entails risks and we have to face them. Between risk and opportunity, I would prefer the latter.
Is ChatGPT the tip of the iceberg as some say? It is the first generalist development that is available to the public in a massive way and it is a tool that will become increasingly powerful as it is trained.. It is true that it has the characteristic that it is generalist and in the law [promoted by the EU] it has been defined that way, although not everyone agrees. We are going to have to get used to this kind of thing, even if it surprises us by its magnitude.
You have commented that we are facing a completely new reality, how do we explain that reality to the average citizen? You have to tell exactly what it is: a technology that is not new but that has reached a point of maturity that makes it seem very revolutionary.. Technically it is more powerful than it was in the past and it is already being applied very widely. We keep it present in one way or another; This is going to create wealth and opportunities for us and it is estimated that world productivity is going to multiply by seven, but at the same time we have to be careful and for that, in the EU we are trying to anticipate to protect ourselves against these risks.. Citizens must be reassured.
In what sense should AI be regulated? There are two ways to deal with it. With an investment and research strategy, and that corresponds to the institutions and we have to have a regulation that looks at the areas in which this AI can be used. What the law does is a risk approach around the use of AI, and establishes different gradations. At the same time, it also establishes a series of obligations in terms of companies or stakeholders providing sufficient information to the entire population about how it works and how it can affect them.
There is talk of four levels of risk, what are the unacceptable risks? It is the use that is ethically undesirable. To give an example, remote biometric identification will be prohibited; uses such as facial recognition that have a significance from the point of view of characteristics such as race or skin color; what is called social scoring, which is a citizen score based on how we behave in society. The latter is something that China does. All these things will be prohibited, as will all those that have to do with massive control over the citizenry.
Does more AI mean less transparency per se? In some cases it may be well justified to think so, but it is precisely about introducing more transparency to know what the decision tract has been to reach conclusions through AI. And what is more important: we have to know when we interact with AI. That is fashionable, for example, with deepfakes (images generated as real when they are not). We have to know that we are facing that or a text that is made with AI, or being attended to in a certain service. It is about reaching that level of transparency and also that people have the right to go to a human person if they want to when making a decision.
We are in times of misinformation and war. Has the context complicated the analysis of AI? Regulation [of the EU] is supposed to be prepared for different scenarios, but it is true that technological evolution is so rapid and so unexpected that it is difficult to make a regulation that anticipates. We try to make the structures sensitive enough to deal with all this.. The law will be adapted in the future, I hope, but the type of content such as false videos or false images are a danger that the law is already going to take into account.
The debate can be faced with optimism, what advantages does AI bring to you? First of all, it is capable of managing and managing data like the human brain cannot.. That does not mean that AI has come to fully emulate the brain, but it is capable of doing so with some things (such as anticipating medical issues).. In the management of water and energy, it will introduce savings that we could not imagine, or with safer transport. It is a great evolution, but it all depends on how we are going to use it. That it be used for the common good is the key, that it does us no harm and that it generates sufficient benefits.
Once the EU puts its own regulation on the table, it will serve as a model for others and to reach international agreements
Does the EU aspire to become the great regulator of AI? Europe is commonplace and yes, we are a great regulatory power and it is our best way to intervene in the world. Our regulation is based on our own needs, we are a democratic space, and that regulation will always be much more protective. That will be good for the EU and for the world.. We are going to be the first and the world is precisely watching us, I can attest to that. Once the EU puts its own regulation on the table, it will serve as a model for others and for reaching international agreements.
And that can generate tensions with other powers? In this who gives first gives twice. We are going to have a competitive advantage, but the rest of the countries are going to be free, of course, to make their own regulations.. If companies want to operate in our space, they will have to do so with our rules and with our principles and values. From my experience, I believe that there is expectation and a certain willingness to reach cooperation spaces with other actors, whether it is the US, with whom we are more united, or with China. Everything will work out.