Osservatorio sul diritto e telecomunicazioni informatiche, a cura del dott. V. Spataro dal 1999, 7795 documenti.

Il dizionario e' stato letto volte.

Segui via: Email - Telegram
  Dal 1999   spieghiamo il diritto di internet  Store  Podcast  Dizionario News alert    


Intelligenza aritificiale 02.10.2020    Pdf    Appunta    Letti    Post successivo  

Dossier sulle responsabilità da Intelligenza artificiale - analisi comparata

Uno studio enorme e imperdibile: Civil liability regime for artificial intelligence European added value assessment.




Potrei citare le 4 pagine di indice ma servirebbe solo a comprendere l'ampiezza dello studio del Parlamento Europeo.

Naturalmente in inglese.

Buona lettura.



The  purpose  of  this  European  added value assessment (EAVA) is  to  provide an evidence-b a sed   evaluation and assessment to accompany the European  Parliament's draft report on a legislative initiative proposal on a civil liability regime for artificial intelligence (2020/2014 INL).

The report has been initiated by the Committee on Legal Affairs (JURI) in accordance with Article 225 of the Treaty on the Functioning of the European Union (TFEU).21.2.Methodology   and   scope   of   the   European   added   value   assessmentThis EAVA focuses on the regulation of civil liability for artificial intelligence (AI) systems a t  E U  l e v el.

There are three main objectives:

  1. first, to analyse the main socio-economic functions of liability rules, and thus, ultimately, explain why liability r egula t ion  s h ou ld be k ey to AI policy;
  2. second, to assess  the current regulation of liability and specifically how the current framework applies or potentially could   apply  to  AI  at EU  and national level; 
  3. third, to assess what could be the potential added  value, measured in quantitative economic terms, of taking common EU a ct io n on a civil liability regime for artificial intelligence.

European added value of EU action on civil liability regime for AI is assessed both quantitatively and qualitatively.

The qualitative  assessment  of  European  added value  is   threefold: first,

Chapter 2 pr o v ides  an  analysis of the  socio-economic functions of liability  rules on the basis of  the literature available; second,

Chapter 3 focuses on EU-level analysis and explains the scope and limits of the  application of the Product Liability Directive to AI; third,

Chapter 4 focuses on the regulation of liability in the Member States, with a specific focus on the applicability of the existing rules to AI systems.  The  main aim of  the  comparative analysis  of current  practices in the Member States is to provide a concise picture of the existing regulatory mixes between fault-based and strict liability rules, and,   more specifically, to map the scope and conditions of the application of the strict liability rules in the Member States.

The question underlying this comparative legal analysis is where the regulation of   AI  should  be   placed  in   the  current  system and  whether  adjustment  is necessary. A b e t ter   understanding of Member States' liability systems can also contribute to discussions at EU level on  possible EU solutions and how they would fit in with existing Member State rules.

The co m p a r ative   legal analysis is based on the original dataset of national rules from 19 EU Member States, s p e c i f i c ally  collected for this EAVA and discussed and presented in full in Annexes I and II.3

The quantitative  assessmentof added value, in Chapter 6, estimates the possible net benefits that could be generated as a result of EU civil law rules on the liability of AI a s  co m p a r ed t o  t h e in d iv idual  actions of Member States. This quantitative assessment is based on a two-step analytical model. First, the added value is quantified, as a net benefit for the EU economy resulting from additional investment  in  research  and   development  related  to  AI;  

second,  the   added   value  is   measured  considering broader economic impacts, including reduced accidents, health and environmental impacts, tax revenues and user impacts.The quantitative assessment builds on the data, methodology and results of the Cost of non-Europe   in  robotics  and  AI:  liability,  insurance  and  risk  management  (2019)  and  European  added va lue assessment: A common EU approach to liability rules and insurance for connected and autonomous vehicles  (2018). 

As will  be  discussed in  more detail in  Chapter 6,  both studies provide only very partial results, as they cover o nly s pecific s ect o r s (i.e. a u t on om ou s  vehicles in the 2018 study) and specific types of impact (mainly research and development). Therefore, the quantification presented in  this EAVA should be considered as preliminary, providing only initial guidance and discussion on the possible costs and benefits of EU civil law  rules on liability of  A

02.10.2020 Europarl

Download Pdf

Segui le novità in materia di Intelligenza aritificiale su via Telegram
oppure via email: (gratis Info privacy)


dallo store:
visita lo store

Dal 1999 il diritto di internet. I testi sono degli autori e di IusOnDemand srl p.iva 04446030969 - diritti riservati - Privacy - Cookie - Condizioni d'uso - in 0.338