New guidelines: EU police office Europol wants to monitor more ethically
Europol has published a framework for ethical technology in law enforcement. It covers topics such as facial recognition and automated decision-making.
(Image: Shutterstock/My Eyes4u)
Europol has investigated the use of technology in law enforcement and developed a "method for ethical decision-making". According to the report published on Friday, the aim of the project is to provide law enforcement agencies in general and their own employees in particular with a structured approach to the use of artificial intelligence (AI) systems and big data analysis, for example. Among other things, the authors identify moral concerns and weigh up the risks and benefits of relevant technical solutions. The aim is to ensure that the use of AI & Co. remains "lawful and proportionate". However, there is already criticism that Europol's powers to mine large amounts of data in particular go far too far.
Specifically, in the first case, the authors examine procedures for analyzing recordings from surveillance cameras. In doing so, they compare the impact on data protection with a potential improvement in public safety. Conclusion: Such analysis methods should only be introduced "for defined use cases and with the approval of a senior official" after public participation. Prerequisite: "The public broadly supports the technology".
There is also an AI risk assessment for cases involving gender-based violence. The focus here is on achieving a balance between the need for predictive tools on the one hand and fairness, accountability and transparency on the other. Observance of these values is a common thread running through the report. Summary here: "The best solution is considered to be proper development with extensive testing and simultaneous use with the current system, as well as transparent public involvement."
Data scraping on the web is not OK
The researchers also scrutinized the implementation of chatbots to prevent the sexual abuse of children on the internet. In doing so, they primarily considered a balance "between proactive intervention and data protection rights". The result: a limited version of such an AI system with a large age threshold between the conversation partners is acceptable.
The experts also focus on the large-scale extraction of data from the web (data scraping). Their conclusion here is clear: as automated investigations regularly violate the terms of use of website operators and platforms, "such scraping of open source data is not acceptable".
The Europol representatives also discuss the ethical and legal concerns of automated online investigations. Here, they propose a middle ground between "everything is permitted and prohibited". Depending on the specific technology, different protective measures are required. In the automated analysis of large and complex data sets, the risks of a creeping expansion of functions must be minimized and a "responsible" use of tools must be ensured.
According to Europol, the case studies are not intended to provide a "final verdict" on the "right" or "wrong" use of a technology. Instead, the aim is to enable law enforcement agencies to "apply relevant methods in different contexts".
Videos by heise
Surveillance powers are hotly disputed
The Police Office sees the report as a "living document" that will be regularly updated with new use cases and ethical considerations. By promoting a structured ethical framework, the agency aims to generally "strengthen public confidence in the use of technology by law enforcement agencies". Europol emphasizes that innovations in police work must "continue to be in line with societal values and legal safeguards".
Thanks to a regulation from 2022, Europol is also allowed to analyze data of unsuspects on a large scale. National law enforcement agencies in particular, such as the German Federal Criminal Police Office (BKA) and the French National Police, have been providing Europol with extensive amounts of data for years. The Hague-based office also helped judicial and law enforcement authorities in Belgium, France and the Netherlands to infiltrate the encrypted communication service Sky ECC. Former EU Data Protection Commissioner Wojciech Wiewiórowski saw Europol's big data powers as weakening fundamental and control rights. He has been feuding with the authority for years.
The #ProtectNotSurveil coalition, which is led by the civil rights organizations Access Now, European Digital Rights (EDRi) and the Equinox Initiative for Racial Justice, is currently complaining that a planned new reform of the Europol regulation would declare "digital war" on migrants. The EU Commission's draft would task the office with leading the fight against the smuggling of asylum seekers and provide it with greater surveillance powers and a significantly larger budget. The alliance sees this as "a systematically flawed proposal that cannot be corrected". The EU Parliament must reject it.
(nie)