Policy Lab 5
27 th April 2023 – Madrid, Spain
The fifth popAI Policy Lab was held online on 27 th April 2023 by the Municipal Police of Madrid, Spain. Approximately 30 participants attended the event, mostly police officers, some of them with experience in the field of AI. The selection of participants was done in a strategic way to ensure a diverse and enriching representation of different profiles related to policing and artificial intelligence. Most of the participants were Municipal Police from different regions of Spain, which provided a broad and varied perspective of the experiences and challenges in the use of AI in the local context.
In addition to the Municipal Police, the event was also atended by a postdoctoral researcher, non-governmental organizations (NGOs) dedicated to the protection of civil rights and privacy, as well as technical advisors from the Madrid City Council with experience in closed-circuit television (CCTV) systems and experts in the field.
During the event, two outstanding use cases were presented: the use of closed-circuit television (CCTV) systems in police activity and the application of drones for the search and location of people. These case studies provided a solid basis for analyzing different aspects related to privacy protection, ethics and accountability, citizen participation and good practices in the training and education of officers.
The methodology used during the event included a combination of case studies, followed by debate and open discussion sessions. The interation and active participation of the atendees was encouraged, who had the opportunity to ask questions, share experiences and contribute their perspectives on the topics discussed.
In addition to the live sessions, links to relevant research were provided so that participants could delve deeper into the topics addressed during the event. Additional communication channels, such as discussion groups and e-mails, were also made available to foster collaboration and information exchange after the event.
Use Case 1:
Use of CCTV in police activity
The first use case was carried out in different phases: A) Presentation of the use case and its importance. B) Discussion of privacy protection in the use of CCTV. C) Ethical considerations and responsibility in handling data captured by CCTV. D) Promotion of citizen participation and transparency in the use of CCTV. E) Best practices for training and education in the use of CCTV.
Briefly, the main points on which the discussion was based and the conclusions drawn are summarized below:
Privacy protection: establishing protocols is crucial to ensure privacy protection in the use of AI. Clearly defining who can access the data and for what specific purposes is essential. Safeguarding personal data when utilizing recorded images is of utmost importance.
Personal and confidential information considerations in the use of AI by law enforcement: it is necessary to defining the appropriate use of personal biometric data based on specific circumstances.
Responsibility for the data: Designating a data controller is important to ensure the proper and responsible use of data. Deliberating on the reliability of the police versus the political establishment in managing this task is a key aspect of the discussion.
Importance of the proper use of data: The focus should lie on how the data is used, rather than solely on its acquisition. Applying either restrictive or utilitarian philosophies depending on the circumstances is essential. If citizens perceive the misuse of data, there is a sense of privacy loss that needs to be addressed.
Information, Transparency and Trust: Inform and be transparent about the proper use of data. Conduct citizen information campaigns on the use of AI. Establish a relationship of trust with society for the use of AI. Citizens must have confidence in police judgment. Raise public awareness of the evolution of the perception of privacy.
Importance of human supervision: There must always be someone controlling the AI system. Police should use AI systems and ensure their proper use.
Use of AI in CCTV systems and participation of society in privacy: Debate on whether the use of AI in CCTV systems constitutes a loss of privacy. What a loss of privacy entails? Society must decide what it is willing to lose in terms of privacy to ensure its security. Clearly establish what constitutes privacy in public spaces.
Legislation and protocols: Justify the purposes for the use of AI. The use of CCTV systems by police forces is well legislated and protocolized.
Levels of data protection: Establish levels of protection to ensure proper use. Custody of data obtained by CCTV.
Bias in the development of the systems: Considering the insights and expertise of law enforcement professionals is essential to address any biases that may arise during the development of algorithms, ensuring fairness and minimizing discriminatory outcomes.
Use Case 2:
Use of drones to search for people
The second use case was carried out in different phases: A) Presentation of the use case and its importance. B) Discussion of privacy protection in the use of drones. C) Ethical considerations and responsibility in drone data collection. D) Encouraging citizen participation in the search for people using drones. E) Best practices for training and education in the use of drones.
Briefly, the main points on which the discussion was based and the conclusions drawn are summarized below:
Comparison between the use of drones and CCTV: Drones are similar to CCTV cameras in terms of privacy concerns, but their mobility and versatility increase the related challenges. Citizen demand for drone systems in security.
Creation of platforms or commissions for the use of AI in security: It is proposed to establish platforms or commissions that include civilian, police, technical and political representatives to regulate and supervise the use of artificial intelligence in security.
Proportionality in the use of AI: The need to establish proportionality between the use of artificial intelligence and the reason for which it is to be used arises. As in the use of police force, the aim is to establish a scale of proportionality in the use of AI.
Informing society: It highlights that technology has impacted the life of society and emphasizes the importance of informing society about the use of artificial intelligence at all levels of security.
Participation of police officers in the development of algorithms: Police officers want reliable and dependable AI. It is argued that police officers should be involved in the development of the algorithms, contributing their police expertise to determine what is to be detected and improve the reliability of the AI. It is highlighted that by leveraging police expertise, AI becomes more reliable. The importance of training and education in relation to artificial intelligence is mentioned.
This part of the debate highlights the similarity in privacy concerns between the use of drones and CCTV, citizen demand for drone systems in security, the proposal to create platforms or commissions to regulate the use of AI, the need to establish proportionality in its application, inform society about its use, and the importance of police participation in the development of algorithms and their training in AI.
In the end, some key points were identified that need to be addressed:
- Information and transparency.
- The citizen must be informed about the use of AI and drones.
- Importance of transparency in the process.
- Society must be involved in the decision on the use of drones.
- Lack of training and knowledge.
- Dichotomy between the technical part and the training of police officers.
- Need to combine both perspectives.
- To orient training towards ethics.
In conclusion, the importance of protecting privacy and guaranteeing the protection of personal data has been highlighted, establishing clear protocols and adequate security levels. Likewise, the need to have human supervision and establish a relationship of trust with society, promoting transparency and providing clear information on the use of AI has been emphasized.
The participation of society in decision-making on the use of artificial intelligence has been a recurring theme, highlighting the importance of creating platforms or commissions that integrate civilian, police, technical and political representatives.
The need to establish proportionality in the use of AI, as in the use of police force, and to inform society about the applications and benefits of this technology has been stressed.
Finally, the fundamental role of police officers’ education and training in the use of AI has been highlighted, taking advantage of their expertise to contribute to the development of reliable algorithms.