Artificial Intelligence: do risks exist?

Artificial Intelligence: do risks exist?

The participants within the session, devoted to Privacy Risk Management in AI, mainly discussed the methodology for identifying actual, but not abstract risks.

Artificial Intelligence: do risks exist?

The session provided such speakers as Dalbir Singh, a representative of Office of the Privacy Commissioner of Canada, Ethan Plato, a member of Office of the Information and Privacy Commissioner of British Columbia, Kristina Zenner, a federal representative on data protection in Germany (Der Bundesbeauftragte für den Datenschutz und die Informationsfreiheit), Roberto Lattanzi, a member of Office on Personal Data Protection (Garante per la Protezione dei Dati Persona) and Sophia Ignatidou, a member of the Information Commissioner's Office.

The speakers emphasized that today there are lots of approaches to assess risks of sensitive data processing, but none of them is actually versatile and integrated, since all familiar approaches have been developed by experts for particular tasks depending on the current regulation in a country. 

In this respect, both a collective and international integration of efforts to minimize the risks from AI system implementation for working with secured data, especially in case of international data transfer, seems to be an extremely difficult challenge. In the course of this session some participants called for applying an interdisciplinary approach to risk assessment at various levels (technological, humanitarian, in the context of human rights and etc.) Moreover, in order to improve the quality of national risk management, some speakers recommended to gear up for legal consolidation of the personal responsibility for actors involved in this process (developers, expert community, representatives of regulatory bodies, human rights activists). It was outlined that a comprehensive risk assessment can only be carried out with the participation of all stakeholders, otherwise faults and underestimation of possible negative consequences in the private life of citizens due to AI integration are inevitable.  

Yury Lindre, an expert on international legal issues of the Competence Center for Global IT Cooperation", noted that the session had revealed significant differences in approaches to risk management, which was not surprising even for leading countries in the field of AI, who were still searching for the optimal formula in regulating complex algorithmic systems. “The only thing that all the speakers somehow agreed on was the need for a comprehensive risk assessment at all possible levels, taking into account the interests of each party involved. 

There is no doubt that the risk management theme will continue to be discussed and updated at the IGF in future as far as national regulations in the AI field are established," the expert emphasized.

Other news