Data protection in AI-powered video surveillance
Video surveillance is now an indispensable part of modern security systems. However, it is often associated with significant data protection issues that affect the privacy of the monitored individuals and raise important legal and ethical questions. Artificial intelligence (AI) offers crucial advantages in this regard. Data protection in AI-supported video surveillance is significantly improved and also makes surveillance much more efficient through real-time analysis. Recent data protection breaches by some Chinese providers such as Hikvision and Dahua, however, highlight the importance of careful provider selection.
Traditional video surveillance systems are widespread and provide basic security, but they have significant limitations and privacy risks. These systems typically rely on two main methods: passive monitoring and active monitoring. Both methods have specific weaknesses that affect both efficiency and privacy
Passives Monitoring
In passive monitoring, systems continuously record video streams so they can be reviewed later. This method requires a lot of time and storage space. Security personnel search through hours of footage to find relevant information, which is inefficient and can overlook important details. The permanent storage of video data increases the risk of hacking and unauthorized access, compromising the privacy of the monitored individuals
Active monitoring
Active monitoring means that security personnel observe the video streams in real-time. This approach is labor-intensive and incurs high costs. The need for shift work further increases operating costs. Additionally, attention levels decrease over longer periods, which can result in important events being overlooked. From a data protection perspective, active monitoring is also problematic. Continuous surveillance by security personnel leads to a feeling of constant monitoring among the affected individuals, significantly impacting their privacy. Moreover, the monitored individuals often do not know who is observing them and for what purpose, leading to a lack of transparency and fairness. Often, there are insufficient safeguards to ensure that the data is used only for the intended purpose, increasing the likelihood of data protection violations. The right to informational self-determination of the monitored individuals is undermined, as they have no control over how and when they are observed
The solution is AI video analysis, but a careful selection of the provider is important
Given the significant limitations and privacy risks of traditional video surveillance systems, both in passive and active monitoring, AI-based video analysis is gaining increasing importance. These modern systems offer a variety of advantages that improve both the efficiency of surveillance and the protection of privacy. AI video analysis uses advanced algorithms and machine learning to process and analyze video data in real-time. This allows them to overcome many of the weaknesses of conventional surveillance methods
However, AI-based video surveillance systems also pose challenges, particularly regarding data protection and security. In recent years, there have been significant data protection concerns and breaches related to video surveillance technologies, especially with some Chinese providers like Hikvision and Dahua
Hikvision
The company has been repeatedly criticized for privacy and security violations. The US government has placed Hikvision on the Entity List, which means that US companies need special licenses to do business with Hikvision. This happened due to the company’s role in human rights violations against Muslim minorities in China, particularly the Uyghurs. Additionally, critical security vulnerabilities were discovered in Hikvision cameras, which allowed attackers to access the devices
Dahua
Dahua was similarly sanctioned. The US government banned the use of Dahua equipment in public facilities due to national security concerns. These concerns relate to potential backdoors in the devices that could be exploited by the Chinese government
These examples show that despite the technological advancements and benefits of AI-based video analysis, data protection and security must not be neglected. It is essential to carefully select providers and pay attention to their data protection and security practices to ensure the integrity of surveillance systems and protect the privacy of the monitored individuals
Important considerations to ensure data protection in AI-powered video surveillance
When selecting the right AI video analysis provider, several important aspects should be considered. First, data protection compliance is crucial. The provider must ensure that it complies with all relevant data protection laws and regulations, such as the General Data Protection Regulation (GDPR) in Europe. This includes transparency in data collection and processing, as well as measures to secure the data. Equally important are high security standards
Transparency and traceability also play a central role. A trustworthy provider should provide detailed audit logs and access controls to ensure that only authorized personnel have access to the data. Providers that use technologies to anonymize individuals and license plates in the video data significantly reduce the risk of data protection breaches
Finally, the ability for local data storage is another important factor. Providers that enable local data storage minimize the risk of data loss and misuse. All these considerations help to find the right AI video analysis provider that meets both security and data protection requirements
AI-powered video surveillance with Vaido: Efficiency and data protection
An outstanding example of such a solution is AI video analysis with Vaidio, which enables efficient and secure video analysis. The AI video surveillance we offer with Vaidio addresses these weaknesses and provides a more privacy-friendly alternative. The system includes both edge-based and cloud-based components, which together offer a flexible and efficient security solution
Edge-based AI video surveillance
- Local data processing: In edge-based video surveillance, the video data is processed directly on-site on a special appliance (hardware). The analysis thus takes place where the data is generated. This significantly reduces latency times, as the data does not need to be sent to a remote cloud
- Local data storage: The processed data is also stored locally. This minimizes the risk of data breaches as the sensitive data is not transmitted over the Internet
Cloud-based AI video surveillance
- Cloud-based data analysis: In cloud-based video surveillance, the video stream is continuously and encryptedly sent to the cloud for analysis, but not stored
- Local data storage: Although the analysis takes place in the cloud, the data is stored locally. This ensures that control over sensitive data remains with the user and minimizes the risk of data loss
Both solutions, both the edge-based and the cloud-based, enable the following functions
- Real-time analysis: The video analysis is performed in real-time, allowing for immediate responses to detected events and eliminating the need for continuous human monitoring. This ensures quick and efficient handling of security-related incidents
- Anonymization: People and license plates can be anonymized to further enhance data protection. This anonymization takes place before further processing or storage of the data. This protects the privacy of the individuals concerned and meets legal data protection requirements
- Event-based monitoring: In areas sensitive to data protection, in addition to anonymizing individuals and license plates or blacking out certain areas, you also have the option to use event-based video surveillance. With this method, only security-relevant incidents are recorded, while all other recordings are not captured. This limits data storage to the essentials
- Access control and audit logs: Strict access controls and detailed audit logs ensure transparent and secure management of video data. This enables traceability of all accesses and changes to the data, which increases the security and integrity of the data
Conclusion
Conventional video surveillance is fraught with significant privacy issues that jeopardize the privacy of the individuals involved and raise legal and ethical questions. The integration of AI and advanced technologies offers a promising solution to address these problems. Through real-time analysis, anonymization, and strict access controls, modern surveillance systems can increase efficiency while protecting privacy. Additionally, legal and organizational measures should be taken to ensure the transparency and security of surveillance systems. The careful selection of providers that meet these standards is crucial to avoid data protection breaches and ensure security.
Anne-Katrin Michelmann
Date 01.09.2024