Background

The AI for Safer Children initiative touches upon what can be considered highly sensitive subjects in its aim to raise awareness and spread knowledge about how AI can support law enforcement worldwide to combat child sexual exploitation and abuse, as well as build the capacity of law enforcement in all UN Member States to implement existing AI tools responsibly to this end. Both the use of new and complex technologies such as AI in the law enforcement context and child sexual exploitation and abuse – a pervasive yet largely unreported violation of children’s rights – are prominent and sensitive topics that demand full consideration of their technical, legal, ethical and societal complexity; a complexity which substantially increases when looked at from a global perspective. 

As an initiative under the auspices of the United Nations, the AI for Safer Children initiative needs to be guided by and fully compliant with the values and principles enshrined in the Charter of the United Nations, including one of the core purposes of the United Nations established in Article 1 that the organization promotes and encourages respect for human rights and fundamental freedoms.

These reasons lead UNICRI and the Ministry of Interior of the United Arab Emirates to establish an ethical and legal process for the AI for Safer Children initiative dedicated to addressing such considerations. 

In the context of this process, extensive research and consultations were undertaken to prepare the AI for Safer Children Ethical and Legal Report summarizing both the process and the measures taken to address the considerations which arose (available here). The first draft of this report was discussed during an Ethics Expert Meeting that took place on 7 December 2021, where ethics experts and other relevant participants provided inputs and analysis. Other relevant experts and stakeholders, including technology providers, were further consulted in this regard.

What does the ethical and legal process consist of?

The ethical and legal process of the AI for Safer Children initiative consists of defining a set of principles that guide UNICRI and the Ministry of Interior of the United Arab Emirates in the implementation of the initiative; identifying the ethical and legal considerations that result from these principles; and designing and implementing the measures necessary to address them to the extent allowed by the scope of the AI for Safer Children initiative.

Although the term “ethics” is often used broadly to encompass legal considerations as well, such as those emergent from human rights law, the process has been entitled “ethical and legal” to highlight that there are also legal obligations at stake such as those deriving from the Charter of the United Nations and international human rights law.

​​​​​​​Click here to read the Core Principles of the AI for Safer Children initiative.

What measures have resulted from the ethical and legal process?

These are some concrete measures resulting from the ethical and legal process of the AI for Safer Children initiative​:

• Terms of Engagement for Law Enforcement Users
• Terms of Engagement for Technology Providers
• Data Protection Policy​​​​​​​
• Learning centre materials on human rights-centred and responsible AI

With a view to foster the responsible development of AI as one of the core principles of the initiative, all the technology providers are asked to leave information about the ethical and legal aspects of the development of their tools, which can be found on each tool’s details page. These include, for instance, possible biases in the training data or whether an ethical, human rights or risk impact assessment has been conducted. The AI Tools Catalogue is constantly revised as we receive updates from technology providers.

Furthermore, the Learning Centre’s Responsible AI section details an AI Toolkit containing seven practical resources available to law enforcement to support them in responsibly procuring, developing or maintaining AI tools for their agency. These resources include a Risk Assessment Questionnaire, workbook and Organizational Roadmap, among others.

What are the next steps?

The ethical and legal process of the AI for Safer Children initiative is an ongoing process, which will be maintained throughout the initiative’s implementation to ensure that any new considerations are addressed as the initiative evolves. Furthermore, not all measures resulting from the ethical and legal process are one-off solutions. Some of these measures may include the establishment of ongoing mechanisms and procedures to address certain ethical issues or identify possible future issues which could be addressed in advance.