By akademiotoelektronik, 15/01/2023

A report advises the European Union on its bill on artificial intelligence

(Etx Daily Up) - The European Union is in full reflection on contours and regulations to be adopted to supervise the use of artificial intelligence via its bill "The artificial intelligence act".A burning and societal subject in particular because of the biometric identification techniques which raise as many ethical as legal issues.

Biometric techniques ask as many ethical as legal questions.A study commissioned by the European Parliament and the Jori and Petit commissions aims to find measures to answer these questions.By definition, biometrics brings together all IT techniques to identify an individual.We can use the physical, biological and even behavioral characteristics of a person to create their profile.

Biometric recognition techniques are today largely associated by the population with mass monitoring, profiling or even decision -making by algorithm explains the report.They are already used everywhere, but lacks a legislative framework.In their report, researchers recommend that they completely prohibit the complete surveillance of natural persons in their privacy environment and to best respect individual rights.

At the heart of most criticisms is a debate on the types of AI considered at high risk."The artificial intelligence act" aims to protect the fundamental rights of citizens of the European Union and to guarantee legal certainty on artificial intelligence.In the absence of this regulation, individuals may not fully realize the impact of this technology in their lives.AI systems can in particular lead to errors, false arrests, algorithmic biases in particular for marginalized groups of skin color, religion or sex.

Enrollment phase and control risk

The most complex question returns to the "enrollment" phase, that is to say the moment when we save and stores a unique model capable of identifying a person among millions.By transforming the physical data (size, eye color, hair color) of a human being into data, this can create major safety problems.If a hacker managed to take control of it, he could trace and recognize an individual anywhere in the world.

The AI Act is one of the first major global political initiatives aimed at protecting people from AI that will be called "harmful".More than 300 companies have expressed themselves on European law around artificial intelligence, present last April.Its adoption will make it possible to classify AI systems according to the risks they present.For example, facial recognition in real time would be completely prohibited apart in certain critical cases.Some companies put pressure on the legislator to extend their interests in the sector.

Creating a regulatory and legal framework common to the 27 countries of the European Union is therefore a necessity and could make it possible to compete with China and the United States.(According to the European Commission, the United States has strongly participated in meetings to comment on the project).The project notably offers the creation of a public database on high -risk AI.The companies concerned believe that the project will cost them too much and will slow innovation in the sector.On the other hand, human rights and ethics defense groups find that the AI law does not go far enough to protect citizens.There is no doubt that the European Union bill may soon return to public debate during the vote of this law by the European Parliament.