CCTV cameras in stores: How to respect the privacy of customers whilst ensuring their security

Video surveillance forms part of the security system in countless stores, whether they’re a small, local store or a supermarket. Yet, how can we reconcile safety with respecting customers’ privacy? We’ll focus in on the points to respect when using video surveillance cameras.

Why install a video surveillance system in a store?

Over recent months, inflation and the hike in the cost of living have led to a rise in shoplifting. For retailers, this crime is the leading cause of inventory shrinkage. Thus, all of these elements have a negative impact on turnover, and retailers are now looking for solutions to combat this issue.

The simplest solution is to install a security camera system. However, depending on the size of the store, you may also need to recruit staff to monitor the screens. This security team will also be tasked with intercepting individuals suspected to be shoplifting. To make sure that these interventions are successful, security agents must be able to rely on evidence, such as video recordings.

Yet, the human eye is not infallible, and there is a risk of misunderstandings or oversights. These reasons demonstrate why the use of artificial intelligence (AI) can be extremely useful in combating and preventing shoplifting.

The implementation of this comprehensive security system is compatible with the privacy of individuals.

Guidelines on the provision of video surveillance cameras in respect of customers’ privacy

The CNIL (French Data Protection Agency) reminds retailers that there are regulations on video surveillance in stores. These are the regulations which apply to places open to the public. There are several guidelines to take into account.

Submitting a request to the prefecture

Before installing any video surveillance cameras in a store, the manager must apply for authorisation for it. There’s a form to complete and send to the regional prefecture or to the Paris police prefecture.

Informing staff and customers of the presence video surveillance cameras

If the retailer has staff, they need to notify the employee representative bodies of its plan to install cameras in the workplace. The CNIL underlines that it is strictly prohibited to use video surveillance cameras to monitor employees’ work.

Customers must be informed of the presence of video surveillance cameras in the store. A sign must be permanently displayed with the following information clearly legible:

• Purpose of the processing and the data conservation period;
• Reminder of customers’ rights in terms of personal data;
• The person to contact in order to exercise their right of access to the images;
• etc.

Places where it is strictly prohibited to install cameras

Retailers are not free to install cameras wherever they want. They are strictly prohibited in changing rooms and in toilets. The CNIL recommends that cameras are position in circulation areas and merchandise areas.

People authorised to view the images

The images must not be freely accessible. Only the store manager and the security staff are able to view the images. The CNIL indicates that you can install a live-stream viewing screen, which can be viewed by anyone, at the entrance of the store.

Retention period for the images recorded

The retention period for images and video recordings must take into account the processing purposes. The CNIL recalls that, in general, a few days is sufficient to carry out verifications. It recommends that the retention period does not exceed one month.

Use of artificial intelligence to detect the gestures associated with shoplifting

This technology is compatible with customer privacy protection, as in no case does it consist of a surveillance system based on facial recognition.

How does the AI integrated in the video surveillance cameras work?

A classic store video surveillance system is essential built around cameras and a video recorder, all controlled and monitored by security agents. To help the agents detect potential shoplifting, the AI analyses the gestures and behaviour of individuals in the store. It therefore doesn’t involve the analysis of customers’ faces to identify them.

In concrete terms, it involves the installation of the Veesion software on the video recorder, which can then be configured according to the case’s specific requirements (24/7 monitoring, etc.). The software can be used with one or more cameras, in line with the system used in the store. An algorithm analyses the video surveillance images in real time. It’s based on three elements:

• The detection of human presence;
• The location of the individual’s limbs;
• The recognition of objects of interest. Objects of interest are those likely to contain concealed items (a caddie, a backpack, a shopping bag or even a handbag).

This technology studies the likelihood of these gestures being associated with attempted shoplifting.

Detection of anomalies for security staff intervention

When the probability of shoplifting (in progress) is high, this constitutes an anomaly. A video alert is then sent to the person in charge of security, who is authorised to intervene and intercept the suspect.

The security agent can, if they deem it necessary, intervene either in the store or when the suspect goes through the checkouts. Thanks to AI, they can, on the one hand, increase their capacity to prevent shoplifting and, on the other hand, react quickly. The video recordings constitute evidence that the agent can present to the person suspected of shoplifting.

Privacy protection

The video alerts sent by the AI to the security staff are automatically deleted after 30 days. This period can be shortened at the request of the retailers using the software.

To reiterate what we’ve previously emphasised, the AI doesn’t analyse customers’ faces. Therefore, it is not possible to recognise individuals who regularly visit the store.

The AI cannot inform the retailer of the identity of someone visiting their store, nor can it implement an identification search using a client record.

To conclude, it is possible to combine both security and respect of customers’ privacy in a store. Nevertheless, the installation of cameras must comply with the applicable regulations and law. The developers of the AI used within the context of retail video protection are well aware of the importance of protecting personal data. In this regard, the use of AI to analyse images is compatible with ensuring the privacy of individuals, as in no case does it allow for them to be identified: this technology only analyses shoplifting gestures, not the faces of individuals.