Privacy and Proportionality in Artificial Intelligence (AI)


Artificial intelligence (AI) is a fast-developing range of technology. There is no unified international definition of AI. The common denominator of various international definitions of the concept is technologies that in some way replicate human cognition in order to process input information to predict an output or reach a certain goal.

The development in processing power and storage capacity have paved the way for new applications of AI. The benefit of AI, compared to traditional programmed analytical methods, is the inherent way AI learns and develops from experience and new data.

The self-developing and complex character of AI together with its emerging use in society makes the technology an important object of legal and interdisciplinary research. AI is emerging in not only private and commercial contexts but also as a support system for decision-making in the public sector in a government to citizen perspective.

The PhD project ‘Privacy and Proportionality in Artificial Intelligence’ examines legal aspects of governmental use of AI-based surveillance in a specific field: monitoring and control of commercial fisheries to prevent and deter illegal fishing.

Fish is Norway’s third largest export commodity and is a renewable, yet not inexhaustible, common resource. Commercial fisheries harvest this resource in a remote location. Research and official reports suggest that violations of fishing regulations could threaten the sustainability of marine ecosystems as well as contribute to an illegal underground economy. In the European Union, electronic monitoring and CCTV is suggested as potential measures to improve the resource control of fisheries. The Dutkat project at the Department of Computer Science aims at limiting the opportunity to fish illegally by combining sensors onboard the vessels with AI to detect the vessels violating fishing regulations.

However, a control regime with electronic surveillance and AI has legal implications. The PhD project focus on data protection and fundamental rights impacts of AI-based surveillance onboard fishing vessels. The overall aim of the Dutkat project is to implement privacy and data protection safeguards in the technology. Such a process is known as privacy by design or data protection by design.

Research Objective

The terms privacy-preservation and surveillance may be regarded by some as a dichotomy. A research objective of the PhD project is to examine whether implementing safeguards into the technology could safeguard the privacy of the people working onboard the vessels in an improved way compared to legal safeguards.

The principle of proportionality is applied in data protection and human rights law. The PhD project also aims to research how different technological solutions in the Dutkat system could alter the proportionality assessments of the system. A specific legal research question in this context is: whether and to what extent the proportionality assessment of the system would vary based on the subsequent governmental use of the outcome of the system.

Research Methods

Human rights conventions and data protection regulations were not drafted with AI in mind. In recent years, there are examples of multiple international initiatives to draft regulatory frameworks of AI. In the European Union, the proposed AI Act is currently negotiated in the European Parliament. The project will use human rights conventions, data protection regulations and AI-specific regulations as the basis for the legal analyses.

Legal research is much dependent on context. The context of the PhD project relates both to the technology, AI, and the specific use case, monitoring of commercial fisheries. A prerequisite for the legal assessment of the privacy and proportionality of the system is context-specific analyses, and the project, therefore, aims to examine both the technology and relevant aspects of the use case.

Summary of the expected or achieved research yield 

The overall aim of the research project is to link legal and computer science research to suggest a prototype of an AI-based monitoring system of commercial fishing vessels that is both effective and protects the privacy of the people working onboard and is a proportionate measure to prevent illegal fishing.



Members:

Bjørn Aslak Juliussen (Principal investigator)