Europol releases report on ethical use of AI video analytics

Video analytics software, which uses CCTVT footage to analyze objects, persons by description, vehicles and more, should be introduced after public engagement and only for defined use cases and with the approval of a senior officer, according to a study released by Europol.

Released last month, the document explores ethical decision-making in law enforcement using technologies such as data analytics and AI.

The use of video analytics software that doesn’t rely on facial recognition has been gaining ground in Europe, most notably during the Paris Summer Olympics 2024. The data analytics was used to analyze video streams for threats in public spaces, including monitoring crowds, locating abandoned objects, detecting weapons and more.

In its report, Europol concludes that police find video analytics useful in supporting serious investigations. In the long term, however, it may also provide a pathway to more intrusive technology such as live facial recognition. The public, on the other hand, may notice more the benefits of CCTV analytics in falling crime rates or if they themselves become a victim of crime. But without proper engagement from the police, deploying video analytics could lead to the erosion of privacy and confidence in police use of AI.

Europol’s report was originally presented in March 2024 by the European Clearing Board for Tools, Methods and Innovations (EuCB). The body was created as an engagement model between EU member states and the Europol Innovation Lab, tasked with exploring how technological innovation can be used in law enforcement while upholding rights.

According to the document, analysis is not meant to provide legal rules. Instead, it offers guidelines for “grey areas” that are open for interpretation.

Aside from video analytics, it delves into other technology use cases, such as model analysis of open-source data scraping, using chatbots to prevent child sexual abuse online, automated analysis of large and complex databases and measuring the risk of reoffending in cases of gender-based violence.

Germany mulls video analytics after violent attacks

Germany is one of the European countries considering introducing more video analytics and even facial recognition features to public spaces following a wave of violent attacks on civilians.

In January, an attack on a group of kindergarten children in Aschaffenburg, Bavaria ignited a debate on better surveillance of people with mental health issues

Senator for Interior for the city of Bremen Ulrich Mäurer argued that protecting the population has priority over data protection and that investigators should be given new powers for facial recognition and AI-fueled data analysis.

“Of course, there is no absolute certainty. It is completely unrealistic that we will record all people. But we are convinced that more can be done in this area,” said the Social Democratic Party (SPD) politician.

The calls were followed by a decision to deploy AI-based video surveillance in public spaces in the German state of Hesse. The technology will be used for data analysis and analyzing video surveillance footage, Hesse’s Minister of the Interior, Roman Poseck announced in February.

“Artificial intelligence can recognize movement patterns, which means it can sound an alarm if there is a blow or a stab,” says Poseck. In special situations, such as terrorist threats, biometric facial recognition may also be used, he added.

The system will be deployed at train stations in Frankfurt in the first half of 2025 with possible expansions in the second part of the year, according to Zeit Online.

Article Topics

biometric identification  |  cctv  |  Europe  |  Europol  |  facial recognition  |  Germany  |  law enforcement  |  video analytics  |  video surveillance