When the European Union said that it was necessary to establish a regulation for artificial intelligenceand stated that his intention was to act exceptionally quickly to respond to this need, we had a fairly clear signal that, if possible, such a standard would see the light of day before the end of this year.
I know that in other areas, talking about a year may seem like a long time, but we must take into account that, when establishing regulations, the European Union must have the opinion of all the countries that make it up, which which supposes a most complex exercise of negotiation and diplomacy, since on many occasions the interests of one state and the other are in conflict, so it is tremendously difficult to find a position that reasonably satisfies the majority, so that in this way it is possible to approve a final document. In other words: in this context, achieving regulation in less than a year is a speed record.
The regulatory process is also quite complex, since consists of several phases, all of them with their deadlines, so that the speed to quickly comply with each one of them is key if, as I said before, it is intended to respond to this new need before the end of the year. And the good news is that, against many forecasts and quite a lot of skepticism about these intentions, today a very important step has been taken in this regard.
As we can read on its official website, The European Parliament is ready to negotiate the first law on artificial intelligence, after the approval of a first draft that begins to shape said regulation. This text, put to a vote in the European institution, has obtained 499 votes in favour, compared to only 28 against and a total of 93 abstentions, which represents a remarkable level of endorsement for a document prepared in such a short period of time, compared to the mean.
The draft contains some uses of artificial intelligence that, if this text goes ahead, will be banned in the European Unionmany of them related to the biometric identification of citizens, the establishment of a classification of high-risk uses, puts in the spotlight the AI systems used to influence voters and the result of the elections, and also establishes a set of obligations for general-purpose systems.
In this sense, those responsible for said models are obliged to assess, and where appropriate mitigate, a plethora of risks (for health, safety, fundamental rights, the rule of law, etc.). They will also have the obligation to register in a registry that will be enabled for this purpose before starting to operate in the European Union, and a particularly hot spot, orforces the generative artificial intelligence models to identify the contents created in this sense.
