Sant Llorenç: Cameras with AI — who sees whom?

Sant Llorenç: Cameras with AI — who sees whom?

In Sant Llorenç de Cardassar, new surveillance cameras have been put into operation at the main access roads. The technology includes automatic number plate recognition (ANPR) and cameras that use artificial intelligence to detect incidents. The municipality justifies this with the aim of supporting the work of the local police and being able to react more quickly to accidents, vandalism or theft. At the same time the administration announced it wants to expand the system to other locations, such as cameras along the Ma-10.

Main question

How can a small municipality like Sant Llorenç use the benefits of modern video surveillance without hollowing out the privacy of residents, employees and holidaymakers?

Critical analysis

Technology alone does not provide better protection. Cameras with number plate recognition see vehicles with German, Spanish or other plates and store metadata; similar systems have been used to count cars at beach access points like Sa Barralina and Es Trenc (sensors and small cameras counting cars). AI algorithms mark movement patterns, count people or recognize 'incident types' — but these assessments are subject to error rates, training data and bias. The crucial questions are: who has access to the raw data? How long are recordings and licence plates stored? On what legal grounds are automated evaluations carried out?

According to a statement, the municipality wants only authorised personnel to process the images. That sounds good but remains vague. There is no public register with technical details (retention periods, encryption, persons responsible for IT security) and no easily accessible statistical reports on how often the system has actually led to deployments. Without this transparency, trust and control are hard to establish.

What is missing from the public debate

Public discussion usually focuses on 'more security' — rarely on the technical and legal limits; see the debate in Palma over AI cameras at Plaza de España and Parc de Ses Estacions. Important points that are hardly discussed: the duration of license plate storage, whether data can be released to police or third parties, the possibility of false alarms by AI, and the obligation to carry out Data Protection Impact Assessments. There is also a lack of clear information on whether sensitive areas (private homes, entrances to clinics, childcare facilities) have been deliberately excluded.

Everyday scene from Sant Llorenç

Imagine the plaça: the bakery on the corner, the small football pitch, the long-time residents sitting together in the morning. A van with German plates stops, the driver collects croissants, chats for five minutes. For the camera this is a data record: plate number, arrival time, duration of stay. For the woman who has lived here for decades it is her shopping. This collision of technology and everyday life shows why questions of purpose limitation and minimisation are so concrete, as debates over cameras at Es Trenc show.

Concrete approaches

Instead of secrecy, the municipality and police should make rules public: an open register of all camera locations with technical details; clear, short retention periods (e.g. 24–72 hours for routine recordings, longer only for ongoing investigations); encryption and logging of all access; regular independent audits of algorithms for bias and error rates; mandatory Data Protection Impact Assessments with summaries published.

Practically: clearly visible notice signs on all cameras, a citizen office where affected persons can request information about stored data, and a small advisory board of municipal councillors, data protection officers and resident representatives that receives a report every six months. As an alternative or addition: more personnel for foot patrols and better lighting at critical points — often simple measures help just as effectively as high tech.

Why it is not an 'all or nothing' issue

Surveillance technology can help solve crimes and coordinate deployments. But it can also enable behavioural control, the monitoring of entire neighbourhoods and data collections that are later used for purposes no one can predict today. Responsibility therefore means: narrowly defined purposes, clear deadlines, independent oversight and public accountability.

Concise conclusion

Sant Llorenç faces a decisive question: Does the municipality really only want to gain security — or does it inadvertently open the door to permanent surveillance? Without transparency and participation of local people, the technology remains a black box. If the administration now discloses how the systems work, who has access and how long data is stored, the camera installation can become a tool for greater security. If not, the village risks exchanging its familiar calm for a permanent gaze.

Anyone drinking their coffee in Sant Llorenç in the morning should know who is watching them — and why.