"At the center is the man, he is the man who asks, this is the European approach", underlined the EU commissioner for digital Maryia Gabriel in presenting the 7 principles at the center of the guidelines, drawn up on the basis of over 500 contributions sent to Brussels by the actors of the various sectors involved. The goal of these 'red lines' is in fact to ensure "trust" in artificial intelligence, which is "essential to benefit" from its use, warned Gabriel.
The first of the principles contained in the EU guidelines on Artificial Intelligence (AI) provides that there must always be human control, because the goal is to improve human action and its rights, not to reduce its autonomy. The second provides that the algorithms must be safe, reliable and resistant to errors or inconsistencies of the different phases of the life cycle of AI systems. The third that i citizens must always be informed of the use of their personal data and have full control of them so that they are not used against them, and this must be done in line with the EU rules on the protection of privacy of the GDPR.
The fourth principle provides transparency, ensuring the traceability of artificial intelligence systems. Fifth, ensure diversity and the non-discrimination, with humans who may be able to modify algorithm decisions taking into account all necessary factors. And with human appeal mechanisms against algorithm decisions, to ensure the responsibility of those who manage the calculation systems in case of damage or accident. In the end, artificial intelligence will have to work for social well-being and environmental, increasing ecological sustainability. "A person must always know when he is in front of a machine and not a human being", explained EU Commissioner Maryia Gabriel, for this reason "AI systems must be recognizable".
A pilot exercise will therefore be launched in the summer in which a wide range of different actors will participate, from non-EU companies and multinationals to public administrations, and will take place under the aegis of the European Alliance for AI and the supervision of the Group of Experts. high-level for AI. The EU will therefore aim to build a coalition at the international level to apply these development principles to artificial intelligence also at the global level, starting with partners who already share this approach, including Canada, Japan and Singapore.
“The guidelines identify key requirements for AI in Europe. They are aligned with IBM's principles of Trust and Transparency and with the concrete actions already taken by the company in the area of Trusted AI. One example is the methods of making AI impartial and explainable, which have been developed in our research labs and are already embedded in AI development platforms (such as Watson OpenScale) that are used by our customers. IBM approves these guidelines and will explore their use within the company and with its customers, and hopes that they can also be followed outside of Europe ", commented Francesca Rossi, member of the group of experts of the European Commission on AI and IBM AI Ethics Global Leader. "
#artificialintelligence #ai #technology #machinelearning #iot #innovation #datascience #tech #robotics #deeplearning #programming #robot #future #coding #internetofthings #business #data #robots #science #electronics #engineering #engineer #code #automation
Source ANSA https://bit.ly/2GbjGxe
