1. Comment
February 26, 2020updated 07 Jul 2020 2:08pm

Computer vision: Key macroeconomic and regulatory CV trends revealed

By GlobalData Thematic Research

As computer vision (CV) technology matures, it will prompt increasing scrutiny from regulators worldwide, especially on the use of facial recognition (FR) technology, which is relatively cheap and easy to implement. In the face of a regulatory vacuum, the lack of open public debate around FR is compromising its legitimacy among the public and raising fears of mass surveillance.

Listed below are the top macroeconomic and regulatory trends in computer vision, as identified by GlobalData.

China

China’s aim is to be the world’s leader in AI by 2030. Chinese facial recognition companies are in a particularly good position, given the government’s efforts to identify and track its 1.4 billion people. China’s social credit system rates each citizen’s trustworthiness and is based on a network of over 200 million surveillance cameras and ubiquitous CV facial recognition technology. The government shares its huge facial images database with Chinese technology companies. This gives them an unparalleled advantage over non-Chinese CV firms. SenseTime and Megvii are the two best-funded start-ups in China’s AI market. They lead the way in facial recognition technology.

Trade wars

China’s efforts to become a dominant player in advanced technologies is regarded by the US as a national security problem. The Trump administration believes the policies enshrined in the Made in China 2025 plan encourage discriminatory treatment of foreign investment, forced technology transfers, intellectual property theft, and cyber espionage. In August 2019, the US imposed 15% tariffs on a variety of Chinese goods, with China reciprocating with new duties. In October the same year, the US Department of Commerce added 28 Chinese companies to a trade blacklist. This banned them from buying products from US companies without government approval. China possesses a number of advantages in AI compared to the US. Despite China’s reliance on the US for semiconductors, the trade war is unlikely to prevent Beijing from pursuing its Made in China 2025 model.

India’s CV data privacy debate

In September 2019 the Indian government announced a facial recognition plan. It invited companies to bid for a system that can take “face images from the CCTV feed and generate alerts if a blacklist match is found”. The police are not allowed to access biometric data from Aadhaar cards. However, facial recognition would allow officers in India’s local police forces to search images through a mobile app. The announcement comes in the midst of an animated debate about data privacy and the handling of personal data.

Proposed EU regulation on facial recognition

The EC is discussing regulation to limit facial recognition. This comes in the wake of revelations about its use for monitoring crowds in areas such as London’s King’s Cross station. The proposed legislation wants to give EU citizens explicit rights over the use of facial recognition data. The collection of biometric data that can be used to identify people is prohibited under GDPR. However, no EU data protection authority has yet enforced the legislation against any company violating it. In the UK the use of facial recognition by the London Metropolitan Police and the South Wales Police has sparked a debate over whether there is any legal basis to use the technology on the general population. A September 2019 High Court ruling that police use of automatic facial recognition technology to search for people in crowds is lawful set an important precedent.

Data privacy

By using software capable of recognising a face in a crowd and scanning large databases of people, CV regularly handles people’s personal data. This poses potential threats to privacy. In particular, the use of facial recognition by the police has raised the biggest concerns. Data protection activists across Europe are looking to reclassify facial recognition data as biometric data. Under GDPR, this is considered sensitive and requires explicit consent to be collected. Similarly, the California Privacy Act includes biometric information within the definition of personal information. The Commercial Facial Recognition Privacy Act is the most recently proposed act in the US regarding privacy on facial recognition. If passed, it would prohibit private entities from collecting and sharing data without consent.

This is an edited extract from the Computer Vision – Thematic Research report produced by GlobalData Thematic Research.