Artificial intelligence is fashionable |Opinion

Futuristic visions apart, no doubt that artificial intelligence (AI) is a strategic booming sector, which directly impacts the process of transformation of practically all areas of the economy.And like any new reality, this also raises new problems and perspectives that need to address.

I do not want to deal with here on more or less futuristic scenarios, but of AI as we know it today.This could be described as the technology that allows programming systems to perform certain tasks based on previous training using large amounts of data from which the "learns" system and generates output information.The Council of Europe refers to artificial intelligence as "a set of sciences, theories and techniques whose purpose is reproduction by a machine of the cognitive skills of a human being"

The legal issues that arise around this technology are very varied, so we will only refer to some of them, starting with biases in their operation.It might seem that the fact that AI is based on mathematical processes, automatically developed, would necessarily lead to objective and non -discriminatory results, but reality is very different.Beyond the induced biases, where appropriate, by the programmers themselves, we cannot forget that technology is based on the use of large amounts of data, from which the machine draws conclusions.A lack of objectivity - of the liberated or not - in the data used for software training will inevitably affect the result.Let's think for example, an artificial intelligence system that was used in the environment of the administration of justice.If the data on convictions that have been used to train the machine collect a majority of sentences for crimes in a certain segment of the population, the machine will tend to conclude that being part of that population segment is indicative of guilt.Taking into account the impact that AI can have on individuals, the first challenge is to ensure that correct and properly varied data are used.

Another area where important challenges are raised is that of civil liability, this point of unquestionable actuality before the reality of autonomous vehicles.Let's think about the difficulties derived from the fact that the suspect of having caused damage - or having participated in it - is an AI system.In such an assumption, which of all the agents involved should be charged with the responsibility?, To the creator of the software, to which the system trained, to which he updated, to the manufacturer who integrated it into the vehicle?In October 2020, the European Parliament issued a resolution with recommendations aimed at the European Commission for the development of a regulation related to civil liability for the operation of AI systems.In it he opts for the objective responsibility for the high -risk and subjective AI systems operator for the other assumptions.We will follow this matter closely.

La inteligencia artificial está de moda | Opinión

Next to the previous ones, another field that raises obvious challenges is that of industrial and intellectual property.There are already numerous examples of works or inventions developed by AI but can you recognize the condition of author or inventor to AI?Our intellectual property system only collects protection through the copyright of the works created by natural persons, although in some cases a legal fiction is applied for those created on initiative and under the coordination of a legal person.With our current legal system, therefore, the authority of AI cannot be recognized.

For its part, the European Office has already rejected several patent requests in which an AI system was designated as an inventor, in the understanding that the legal system requires that the inventor be a natural person and that the rights that are recognized to the inventorYou cannot recognize a technological tool, lacking legal personality.However, outside the EU there have already been some pronouncements in another sense.We will see how this matter evolves.

We will end up pointing out that on April 21, 2021, the European Commission published a regulation proposal by which harmonized norms are established in the field of which, if approved, will involve a transverse regulatory framework that will regulate the introduction into the market, commissioning and use of AI systems, depending on the risk that they imply for the health, security and fundamental rights of the people.This Regulation contemplates, among other matters, issues related to the protection of personal data, so important also in this area;Specifies what techniques of AI are prohibited;regulates the governance of training data, validation and proof;the need for human surveillance for high -risk systems;The demands in cybersecurity, surveillance and monitoring, etc..

The sanctions that are collected are very relevant, so you have to closely follow the evolution of this standard.

Siete beneficios de la CPP y la CPI: resultados que los acreditan en EuropaLa UE contra el arbitraje de inversiónAtentos al tipo mínimo mundial en el Impuesto sobre es