UNICRI’s programmes aim to promote national self-reliance and the development of institutional capabilities. To this end, UNICRI provides a one-stop facility offering high-level expertise in crime prevention and criminal justice problems. Technical co-operation is enhanced by the use of action-oriented research to assist in the formulation of improved policies and concrete intervention programmes. Institutional and on-the-job training of specialized personnel form an integral part of UNICRI activities.
Description of Activities on AI
Through its specialized Centre for AI and Robotics in The Hague, UNICRI advances understanding on the risks and benefits of AI, robotics and related technologies vis-à-vis crime, terrorism and other threats to security and seeks to support Member States to leverage the potential of these technologies in a responsible manner.
Relying on its convening power as a UN entity, UNICRI organizes several events, including: information-sharing symposiums, technical workshops, training courses, multi-stakeholder policy discussions, and high-level awareness-raising and visibility events. UNICRI is also exploring the conceptual design and development of AI-based tools to, for instance, prevent, detect and facilitate the prosecution of the perpetrators behind online child sexual abuse material and to interpret irregularities in financial transactions that might indicate the financing of terrorism.
The Centre was established by UNICRI in September 2017 with the support of the Municipality of the Hague, the Ministry of Foreign Affairs of the Kingdom of the Netherlands, the Ministry of Interior and Kingdom Relations and strategic partners from the private sector.
Project 1: INTERPOL-UNICRI Global Meetings on AI and Law Enforcement
July 2-4, 2020. Participants from law enforcement agencies of UN and INTERPOL Member States gathered for the second annual INTERPOL-UNICRI Global Meeting on AI for Law Enforcement to discuss AI use cases and technology domains and examine their relevance for law enforcement. The ethical, legal and social implications of the use of AI by law enforcement were also discussed as a central theme of the meeting and the development of a responsible AI innovation toolkit for law enforcement was identified by UNICRI and INTERPOL as a recommended follow-up action.
Project 2: Launch of UNICRI’s Centre for Artificial Intelligence and Robotics
July 10, 2019. Ambassadors from The Hague-based Embassies, representatives of international organizations and eminent partners from academia and the private sector gathered at the Peace Palace in The Hague to formally launch the UNICRI Centre for AI and Robotics. The event was organized by the Ministry of Foreign Affairs of the Netherlands, the Municipality of The Hague and UNICRI.
Project 3: Technical Workshop: Deepfakes and Video Manipulations
November 11, 2019. In June 2019, UNICRI presented a challenge on deepfakes at the Hackathon for Peace, Justice and Security. The purpose of this technical workshop is to keep up the momentum after the hackathon and foster further interest in the topic of detecting video manipulations, as well as to further explore avenues for continuing to work on winning solution designed and built at the hackathon. The workshop looked at the current status and prevalence of video manipulations – to understand how major international organizations such as the UN, Europol and NATO are tackling the risks and improper use of video manipulations– and explored the technology behind video manipulation detection, and what we can and should do to make sure video manipulations can be detected and not used improperly.
Project 4: Training: Artificial Intelligence and the Judiciary: AI Technology Today and Beyond
October 15-16, 2019. This training course, organized in cooperation with the Dubai Judicial Institute and the Government of Dubai, sought to enhance judicial knowledge on artificial intelligence. The course focused on issues such as the legal definition of AI, the dangers and realities of algorithmic and output bias, the attribution of responsibility in cases where harm is done by an AI system, and specific AI applications that can support the judiciary in the performance of its duties. The course was attended by more than 100 participants from the UAE, as well as by representatives from 12 members of the Euro-Arab Judicial Network.
Project 5: Roundtable: Decoding Biases in AI: Does AI have a Diversity Problem?
January 21, 2020. Bringing together academia, policymakers and the private sector, this roundtable discussed the social impact of artificial intelligence, diversity in the tech sector and whether it is possible to build fair algorithms. The roundtable was organized by the T.M.C. Asser Institute, the Embassy of Switzerland to the Kingdom of the Netherlands and UNICRI in coordination with the International Gender Champions Den Haag.
Project 6: Panel Discussion at Davos: The Role AI can Play in Stopping the Online Sexual Abuse of Children
January 23, 2020. UNICRI, the Bracket Foundation and Interfaith Alliance for Safer Communities (IAFSC) organized a panel discussion at Davos on the role technology can play in stopping the online sexual abuse of children organized. Questions regarding the potential of advances in technology to reduce abuses such as the online spreading of child sexual abuse and exploitation material were discussed, as well what kind of new tools law enforcement could rely on to prevent and counter these crimes through the identification of victims and perpetrators.
Project 7: Virtual Discussion Room on COVID-19: How Can AI Support Law Enforcement?
May 13, 2020. INTERPOL and UNICRI organized a Virtual Discussion Room focused on the role AI can play in supporting law enforcement to preserve public safety and social order during the COVID-19 pandemic, as well as contain the spread of the virus. During this event the second report on AI for law enforcement “towards responsible AI innovation” was launched.
Challenges and Opportunities
Ensuring human rights complaint AI and building public trust in the use of AI by law enforcement have been and continue to be the primary challenges UNICRI faces in terms of the use of this technology for crime prevention. This is relevant now more than ever in light of numerous recent developments, including privacy law violations against private entities supplying facial recognition software to law enforcement, concerns regarding accuracy and racial discrimination in the use of facial recognition software, racial equality movements following the death of George Floyd in the United States and the related anti-policing sentiment. A further challenge UNICRI has faced in terms of exploring the development of practical AI applications for crime prevention as part of pilot projects concerns access to data. The data required is often sensitive in nature due to both privacy and security concerns, which creates legal hurdles. Often data is linked to ongoing investigations which can equally limit access.
As criminal investigations become increasingly data heavy and as law enforcement are being asked to maintain operations, or even do more, with less, AI can be a powerful tool in supporting law enforcement to continue to prevent and control crime if the appropriate conditions are in place. There are numerous unexplored opportunities for employing AI to prevent crimes.