UNICRI: Organizational AI readiness - key for Azerbaijan - INTERVIEW

Innovations in the field of artificial intelligence (AI) open up new prospects for states, but at the same time raise complex questions - from ethics and human rights protection to the organizational readiness of institutions. Azerbaijan recently presented its national AI strategy for 2025-2028, focusing on creating a safe and sustainable ecosystem.

In an interview with Report, Odhran McCarthy, New York liaison officer/program officer (Artificial Intelligence) at the UN Interregional Crime and Justice Research Institute, told how UNICRI's approach aligns with this strategy, what opportunities for cooperation are emerging, and why AI literacy is becoming a key factor for the future.

Report presents the interview:

- Azerbaijan recently presented its artificial intelligence (AI) strategy for 2025-2028, aimed at creating an innovative and safe AI ecosystem. How does UNICRI's approach to AI governance relate to Azerbaijan's national strategy and what opportunities exist for cooperation?

- I think for Azerbaijan, a very important emphasis here is placed specifically on organizational readiness, on preparation for this. And in many ways, this is much less related to the technical side than to the human factor. How to create understanding within an agency about what AI is? How to develop potentials, culture, expertise, specialists and so on? I think this is exactly what is directly applicable to Azerbaijan and most likely aligns well with its own strategies and vision, and, frankly speaking, with any countries that are preparing for innovations and seeking to implement them responsibly.

- UNICRI collaborates with Interpol in developing tools for responsible innovation in artificial intelligence in law enforcement agencies. Given Azerbaijan's desire to modernize its law enforcement agencies, which specific aspects of these tools might be most useful for Azerbaijani police?

- When creating these tools, which UNICRI and Interpol developed with the support of the European Union, we primarily tried to adapt what is happening in the AI field to the needs of the law enforcement community.

It is obvious that AI is something new for all of us, and we are all trying to keep up with its development. Therefore, in many ways, the task of the toolkit was to provide fundamental knowledge and understanding of AI specifically in the context of law enforcement, to adapt the description and explanations to their specifics, to make the material accessible to them. In this regard, I believe that the greatest value of the toolkit is the introductory guidance materials: what AI is from a non-professional perspective, how it is applied in law enforcement, what is the role of human rights, ethics, and what "responsible AI" means for the police. All this is basic introductory information presented in their language, and this, in my opinion, is the most valuable contribution for this community or for any agency that is preparing to work in this field. - Azerbaijan seeks to strengthen international cooperation in AI research and development. How can UNICRI help establish closer ties between Azerbaijan and other countries or international organizations in the field of AI, especially in crime prevention and justice? - This is indeed a new area, and we are all just beginning to figure out how to implement innovations and how to do it responsibly. And I think we all have something to learn from each other, and this is a big part of our approach. As you mentioned, we have developed a toolkit, and the next stage of our work is to implement it together with agencies around the world. This work is built on cooperation. As part of the new phase of the project, called AIPOL, funded by the European Union and implemented jointly with Interpol, we have gathered five agencies from around the world so they can exchange experiences and discuss what responsible AI implementation means for their organizations. How to create processes and structures that will allow innovations to be implemented responsibly? Collaboration and working together is key to our approach. And hopefully, in future phases of this work, we will have the opportunity to involve Azerbaijan as well. - As AI technologies develop, concerns about privacy and ethical aspects are growing. What measures does UNICRI recommend to ensure that AI developments in Azerbaijan comply with international human rights standards and ethical principles? - Our toolkit emphasizes precisely the concept of organizational readiness - and this is indeed a key point. Before starting to use AI or integrating any tool, it is necessary to conduct a self-assessment of the agency: discuss it with the legal department, human resources, ethics office, public information office; understand the current capabilities of the organization, identify gaps and weaknesses, analyze the internal culture of the agency. This exact early and honest self-assessment is, in our view, the first crucial step to reduce future risks and problems that will inevitably arise during technology implementation. Thus, self-assessment and reflection at early stages help prepare for more complex challenges related to AI integration in the future.

- Given the importance of AI education, how can UNICRI support Azerbaijan in developing local expertise and capacity in the areas of AI ethics, governance, and its application in crime fighting and justice sectors? - Yes, in my opinion, the issue of "AI literacy" is now one of the most relevant. We need to expand our own knowledge and understanding of how AI affects us - as a society, as individuals, and as organizations. In our work with the law enforcement community, we strive to develop this AI literacy. Already in November, we are starting the first training course for law enforcement leaders - this is a kind of AI literacy program for their senior leadership. The goal is for them to begin discussing what AI is, how it affects their work, what they need to know. It's a sort of "AI 101." And this is a direction that we are actively developing, and I believe that such initiatives are needed much more broadly, not just in the law enforcement sector. Improving literacy in the field of AI is, without doubt, a key step for the near future.