Towards Responsible AI Innovation: Second Interpol-UNICRI Report on Artificial Intelligence for Law Enforcement

‘Towards Responsible AI Innovation’ report on AI for law enforcement was published on May 19, 2020. This new report is the most recent product of the collaboration on AI between the Innovation Centre of the International Criminal Police Organization (INTERPOL) and the United Nations Interregional Crime and Justice Research Institute’s (UNICRI) Centre for AI and Robotics.

INTERPOL’s Innovation Centre and UNICRI first collaboration was held in Singapore in July 2018 in the framework of the first Global Meeting on AI for Law Enforcement. Following these meetings, a first report was published. It included a series of recommendations for the future use of AI and robotics in law enforcement, as well as specific suggestions for police chiefs. The second edition was scheduled to dive deeper into the ethically responsible integration of AI and machine learning into law enforcement to capture and evaluate data and on the practical experiments and projects being implemented. The second Global Meeting on AI for Law Enforcement took place in Singapore in July 2019.

This new publication underlines the need for “additional guidance and support by the law enforcement community to facilitate its adoption of AI and, in doing so, to avoid, not only the ‘possible’ but the ‘inevitable’, pitfalls of its use’’.

‘’Through the close cooperation between INTERPOL and UNICRI, we hope to ultimately contribute to filling this gap, by supporting the identification of current and potential use cases and by providing guidance for the development, deployment and use of AI systems in law enforcement in both a lawful and trustworthy manner’’

The report captures the essence of discussions held over the 2019 meeting and complements it with further analysis and insights on recent developments of relevance regarding the use of AI for law enforcement, including current trends in AI domains and AI regulations.

For serving as a reference to law enforcement agencies, the report is divided into seven sections:

  1. Chapter 1 focuses on the continued expansion of AI for law enforcement: it describes the general landscape of AI use in law enforcement, noting in particular the growing interest of the law enforcement community and underscoring the need for collaboration between the stakeholders. It also highlights other threats related to advancements in AI, in particular their malicious use by criminals and terrorist groups.
  2. Chapter 2 details the four main technology AI domains considered of relevance for law enforcement by INTERPOL and UNICRI: audio processing, visual processing, resource optimization and natural language processing.
  3. Chapter 3: focuses on different uses cases of AI from law enforcement agencies and aims to further support  law enforcement in conceptualising potential practical applications of AI and to foster the exchange of experiences and lessons learned in Australia, Germany, Japan and Norway.
  4. Chapter 4 deals with the responsible use of AI by law enforcement and recalls the general principles that law enforcement should endeavour. It also underlines the specific legal challenges, considerations on the importance of social acceptance by the public and recent ethical frameworks and statements of principles being developed in Europe.
  5. Chapter 5 focuses on the policing needs and recommended actions for realizing the responsible use of AI for law enforcement.
  6. Chapter 6 offers an operationally oriented toolkit to support and guide law enforcement in the design, development, and deployment of AI in a responsible manner.
  7. Chapter 7 finally defines a roadmap and select points of action for the international law enforcement community, policy-makers and intergovernmental organizations to consider in order to support the development of the toolkit and to further develop and promote the concept of responsible AI for law enforcement.

In conclusion, as ‘’AI is here to stay. The question we must therefore grapple with is not if law enforcement should use AI, rather it is precisely in what ways can or should law enforcement use AI and how it does so in the most responsible and appropriate manner’’.


Like this article?
Share on Facebook
Share on Twitter
Share on Linkdin
Share by Email