English Section
Follow us on Google

Follow us on Google to get our latest news at the top of your search results

Cameras everywhere: How AI is changing surveillance

09.10.2026 14:00
Smart glasses, robotaxis, license plate scanners: There are more and more cameras in public spaces. What is new is that AI can analyze the images en masse – and private companies are making a business out of it.
File photo:
File photo:AFP

A car turns the corner, with a sensor tower on the roof, and a man in a yellow safety vest at the wheel. Since late summer, Google's sister company Waymo has been preparing its robotaxi service in Munich, still with drivers at the helm. By the end of 2027, the operation of self-driving cars is set to begin. Such vehicles constantly capture their surroundings, using Lidar, Radar – and also cameras.

They are not the only ones. Meta sold more than seven million smart glasses by 2025. Since 2023, the glasses with built-in cameras have been available in stores. The fact that they record is indicated only by a small light on the frame.

From images to data

What is new is not only the number of cameras, but also what happens afterward. The days when an investigator had to manually analyze footage from public cameras for hours by continually rewinding and fast-forwarding the video are long gone. Today, AI automatically searches vast amounts of images and chats, in countries like China using facial recognition in real time.

🎧 How is AI changing our lives? And which AI programs are really important in my daily life? Answers to these and other questions are discussed weekly by Gregor Schmalzried, Marie Kilg, and Fritz Espenlaub in "The AI Podcast" – the podcast from BR24 and SWR.

Automatic image recognition is also fundamentally possible in private communication and digital image storage. In countries like the USA, tech companies are required to report abusive imagery on their platforms to the authorities. However, searching is voluntary. In the EU, the so-called Chat Control 1.0 was extended until 2028: an exception rule allowing Meta or Google to voluntarily scan unencrypted chats for abusive representations.

Surveillance as a business model

How surveillance in public spaces becomes a business is exemplified by the US company Flock. The company manufactures cameras that automatically scan vehicle license plates on public roads in the USA. More than 5,000 police agencies are customers, and investors value Flock at over eight billion dollars. The technology is, to say the least, controversial: There are repeated reports of police officers abusing the cameras to, for instance, stalk private contacts. In many places, destroying and spraying Flock cameras in the USA has become a kind of popular sport.

Where evasion ends

Resistance is also rising elsewhere: Artists and resourceful entrepreneurs are selling T-shirts with special patterns or reflective makeup that is supposed to protect against automated facial recognition. However, it is not always possible to evade surveillance: For instance, since late 2025, all non-US citizens must have their faces scanned when entering and exiting the USA – only US citizens can object.

The debate about the tension between surveillance and civil rights is not new. The balance between goods such as privacy on one side and the prevention and resolution of crimes on the other is among the major political issues that must be renegotiated repeatedly. To learn more about how technological change influences these debates, tune in to the latest episode of "The AI Podcast."

An article written by Fritz Espenlaub (BR), initially published on 7 October 2026, 16:02 (CEST)