The Era of AI Facial Recognition Technology... 'Big Brother' Concerns in Developed Countries [CCTV Surveillance Society]②

Evolving CCTV Market with Technological Advances
Need for Social Consensus on Usage Purpose and Scope

A scene from a movie where criminals are identified among people walking the streets through CCTV is becoming a reality. This is due to the rapid growth of the CCTV market integrated with artificial intelligence (AI) technology. However, approaches to this vary by country. While China, Malaysia, and others actively use it for investigation and arrest, advanced countries in the US and Europe are cautious due to concerns about privacy invasion. As cutting-edge CCTV systems are being introduced one after another in South Korea, voices are calling for sufficient discussion and social consensus.

Tianwang System of Suzhou Public Security Bureau featured in a Chinese documentary.
[Image source=Yonhap News]

China's 'Tianwang Project'... Effective Investigation, Human Rights Controversy

According to global market research firm Fortune Business Insights on the 19th, the global CCTV market size is expected to grow from $35.5 billion in 2022 to $105.2 billion by 2029. Among this, AI CCTV is projected to account for about 60%.

Facial recognition technology refers to the technology that extracts facial shapes or features such as eyes, nose, mouth, and forehead to recognize faces and verify identities by comparing them with stored database information. Facial images are collected through various methods, including video images captured by CCTV in public places or publicly available photos on the internet. In this regard, facial recognition technology can cause massive human rights violations.

China has been conducting the 'Tianwang' (天網, Sky Net) project since 2015. It is a national project to install advanced CCTV in major cities for crime prevention purposes. Through this, in 2018, authorities succeeded in arresting a wanted criminal who had killed seven people and fled for 20 years. The suspect disguised his face through plastic surgery during the escape period and used fake identification cards. However, Chinese public security identified the suspect using CCTV equipped with big data and facial recognition technology and eventually succeeded in arresting him.

Malaysian police have introduced body cams equipped with facial recognition technology. They quickly search images of suspected individuals and cross-check and verify them against the police database. In this way, some countries have seen facial recognition technology effectively apprehend violent criminals and long-term fugitives.

US and EU Ban Completely... UK Uses It Restrictively

In contrast, advanced countries take a cautious stance on the introduction of facial recognition technology due to concerns about personal information protection and privacy invasion. In 2019, San Francisco, California, enacted an ordinance banning law enforcement agencies from using facial recognition technology. This was due to technical issues where the technology showed a high rate of misidentification for Black people and women. According to research by the MIT Media Lab, facial recognition software correctly identifies white male faces 99% of the time, but the misidentification rate for dark-skinned women’s faces reaches 35%.

The AI Act being promoted by the European Union (EU) is regarded as a comprehensive regulation with the character of a fundamental law. Except for some exceptions, it fundamentally prohibits real-time remote biometric recognition and the construction of facial recognition databases.

The United Kingdom has introduced facial recognition technology restrictively for investigations. Since January 2020, the Metropolitan Police Service has deployed live facial recognition cameras across London to respond to violent crimes. Real-time facial surveillance is conducted only in specific locations where there is a possibility of apprehending criminals. Targets include suspects and wanted persons involved in serious violent crimes, firearm and weapon offenses, and child sexual exploitation allegations.

The system stores only footage where alerts have been triggered for up to 31 days, and facial recognition data of ordinary citizens is automatically and immediately deleted. If a surveillance target is arrested, the footage is preserved until the judicial process concludes. This is evaluated as an appropriate use of facial recognition technology within legal and ethical boundaries.

South Korea provides intelligent CCTV to individuals requiring personal protection. It compares pre-registered facial information of the requester, family members, etc., to identify outsiders and issues warnings in case of abnormal behavior. AI CCTV is also used to find missing persons such as elderly dementia patients and lost children, and to warn about risks in densely populated areas. Under the Personal Information Protection Act, facial feature points extracted by facial recognition technology are classified as sensitive information. Separate legislative measures are required for the police to use facial recognition CCTV in criminal investigations.

Professor Choi Kyung-jin of Gachon University’s Department of Law emphasized, “Facial recognition CCTV is technically feasible even now. The problem is that the state could surveil the entire population,” adding, “A social consensus on the extent of legal allowance regarding usage procedures, purposes, and restricted methods must precede.”