Usklađivanje zahtjeva OUZP-a s izazovima umjetne inteligencije

master's thesis

• Title (english): Aligning the requirements of GDPR with the challenges of artificial intelligence
• Author: Dorian Šegović
• Mentor: Tihomir Katulić (mentor)
• Committee member: Tihomir Katulić (predsjednik povjerenstva)
• Committee member: Hrvoje Lisičar (član povjerenstva)
• Committee member: Marko Jurić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Law; (Information Technology Law and Informatics); Zagreb
• Defense date and country: 2024-07-08, Croatia
• Scientific / art field, discipline and subdiscipline: SOCIAL SCIENCES; Law

Abstract

Zaštita podataka stekla je izrazitu važnost s pojavama novih tehnologija poput umjetne inteligencije (UI). U tom segmentu vrlo je naglašena zaštita temeljnih ljudskih prava poput prava na privatnost i prava na zaštitu osobnih podataka. Kroz to vodi Opća uredba o zaštiti podataka (OUZP), njezin okvir i najrelevantniji članci. OUZP daje zakonodavni okvir kako bi pravilan i pošten postupak obrade podataka trebao izgledati. Pretpostavka je da, iako odredbe OUZP-a ostavljaju prostora za interpretaciju novih tehnologija, brojna pitanja vezana uz zaštitu podataka, a u vezi s umjetnom inteligencijom, nemaju izričit odgovor u OUZP-u. To može dovesti do neizvjesnosti i troškova te može nepotrebno ometati razvoj aplikacija umjetne inteligencije. U radu se analizira kako se obveze voditelja obrade odražavaju na poslovanje organizacije, konkretni učinci ljudskog nadzora automatiziranih odluka UI-ja, načela transparentnosti, minimizacije, zakonitosti, točnosti, nepristranosti, obveze informiranja i ispunjavanja prava ispitanika, pristup utemeljen na riziku, kategoriziranje sustava UI-ja prema predloženom Aktu o umjetnoj inteligenciji i koje su potrebne tehničke i organizacijske mjere poduzete za očuvanje privatnosti od strane organizacija, poput anonimnosti i pseudonimnosti, te korištenje postupaka certificiranja za dokazivanje usklađenosti prema propisima EU-a.

Abstract (english)

Data protection has become extremely important with the emergence of new technologies such as artificial intelligence. In this segment, the protection of fundamental human rights such as the right to privacy and the right to the protection of personal data is highly emphasized. The General Data Protection Regulation, its framework and the most relevant articles guide us through this. The GDPR provides a legislative framework for what a proper and fair data processing procedure should look like. The assumption is that, although the provisions of the GDPR leave room for the interpretation of new technologies, many questions related to data protection and artificial intelligence do not have an explicit answer in the GDPR. This can lead to uncertainty and cost, and can unnecessarily hinder the development of AI applications. The paper analyzes how the processing manager's obligations reflect on the organization's operations, the concrete effects of human supervision of automated AI decisions, the principles of transparency, minimization, legality, accuracy, impartiality, obligations to inform and fulfill the rights of respondents, a risk-based approach, categorization of the AI system according to the proposed Act on artificial intelligence and the necessary technical and organizational measures taken to preserve privacy by organizations such as anonymity and pseudonymity and the use of certification procedures to demonstrate compliance with EU regulations.

• Keywords: Ključne riječi: Opća uredba o zaštiti podataka; umjetna inteligencija; tehničke i organizacijske mjere zaštite podataka
• Keywords (english): General data protection regulation; artificial intelligence; tehnical and organizational data protection measures
• Language: croatian
• URN:NBN: urn:nbn:hr:199:501257
• Study programme: Title: Law; Study programme type: university; Study level: integrated undergraduate and graduate; Academic / professional title: sveučilišni magistar/magistrica prava (sveučilišni magistar/magistrica prava)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2024-07-09 13:10:03