Today, retailers are seeking ways to improve customer satisfaction and compete for market share. Facial recognition can potentially drive the revolution in the retail landscape. It is an emerging technology in biometrics, which works by capturing images, extracting facial features, matching databases and identifying individuals. In the past decade, it has been applied in security and law enforcement, such as border identity checks and criminal investigations.

In the future, will it be possible for us to enter the store and be greeted by name, receive tailored recommendations based on past purchases, and complete payments just with a smile? Facial recognition has influenced the retail industry, offering opportunities for efficiency, security and customer engagement. However, similar to any disruptive innovation, risks and ethical dilemmas also exist and should be addressed through governance strategies.

Opportunities of facial recognition in retail

The use of facial recognition surged in retail stores during the pandemic, making it possible to track customer flow and body temperature with fewer employees and complete contactless payments. Other applications for facial recognition have emerged, such as preventing shoplifting and improving security, presenting opportunities and social value for customers and retailers.

Enhancing customer experience

For customers, it boosts convenience of payment. Without taking out bank cards and waiting in line, customers can be identified and payment can be quickly authorized. It is convenient for those holding a baby or many things in hand. Facial recognition can verify customers’ age when they are purchasing alcohol, functioning as a good combination with self-checkout systems. Moreover, it improves shopping experiences. When customers enter the shop, their basic information such as past purchases and preferences can be generated and transmitted to the staff, thus offering tailored recommendations. Personalized advertisements and loyalty programs can be delivered to consumers accurately.

Cost reduction and security improvement

For retailers, facial recognition reduces costs and enhances security. There were around 300,000 reports of shoplifting in England and Wales in 2022. Facial recognition helps detect identified shoplifters by searching databases and sending security alerts in time, making the store safer and saving money from reduced thefts. For example, the UK’s facial recognition firm Facewatch, provides access to databases and claims its solution can help reduce store theft by 35% in the first year. Additionally, the technology can be integrated with check-in systems and smart door locks, making it more efficient for employee management.

Alignment with Sustainable Development Goals (SDGs)

If properly used, this technology aligns with SDG 8, Decent Work and Economic Growth and SDG 9, Industry, Innovation, and Infrastructure, by improving retail efficiency and productivity and applying advanced technology solutions for retail operations. Integrated into the digital systems, the technology contributes to better products and services with other emerging technologies such as AI and IoT. Moreover, it contributes to SDG 16, Peace, Justice, and Strong Institutions, as it is crucial in deterring criminal activities more precisely to strengthen security and foster societal well-being with technology iteration and AI support.

Image Source: https://heimdalsecurity.com/blog/the-ahrc-calls-for-protecting-human-rights/

Risks and ethical dilemmas

For some, this all may sound great. But to others, this is perhaps rather dystopian. The technology is still young and can cause problems and have unintended risks.

Privacy and data protection challenge

Privacy may be challenged when personal information is quickly generated through facial data. As it is unclear how retailers will process and manage customer data, facial recognition may affect individuals’ reasonable expectations of privacy. People are reluctant to disclose information gathered from facial recognition to retailers due to a lack of trust. Moreover, data may be collected involuntarily as there is sometimes no need for physical interactions and individuals may lose consent rights. If unnecessary facial data is being collected, retailers should ideally delete it. However, this cannot be guaranteed.

Potential for bias and discrimination

Facial recognition algorithms rely on training data to recognize patterns and features. If the data is not representative, it would lead to biases. According to the research, the racial diversity of facial databases impacts how accurate the results are, especially when identifying people of colour. This could result in false accusations and arrests. For example, Apple was sued because the facial recognition system wrongfully identified a student as a shoplifter in 2019.

Human rights and societal issues

Facial recognition cannot solve the fundamental issue of shoplifting. It misidentifies people of colour and LGBTQ+ due to algorithmic bias, who are more likely to suffer from monitoring and harassment. Rights Groups argue that the soaring cost of living, such as basic food and energy costs, is making more people consider shoplifting. Thus, retail stores should respect customers’ rights, and governments should address the root causes of shoplifting, such as the cost-of-living crisis and poverty.

Requirement for governance strategies

Regulatory framework

Regulatory frameworks are important for ethical and equitable use of facial recognition. They should address key concerns such as data privacy, transparency and accountability in different regions and nations.

Under the framework of the General Data Protection Regulation, the EU plays a leading role in data protection. In the proposed regulatory regime for artificial intelligence of the European Commission, facial recognition is strictly prohibited. Countries such as Australia and US have set new privacy and data protection laws. For example, Texas has updated biometric legislation, setting requirements for private and commercial entities.

Therefore, as current data legislation in many regions could not eliminate all the challenges resulting from facial recognition, more strict and comprehensive regulations should be set for both public and private sectors.

Responsible innovation

The framework of responsive innovation should be considered in facial recognition governance. It includes four dimensions – anticipation, reflexivity, inclusion, and responsiveness (Figure 1). For ‘anticipation’, governance efforts are made to anticipate privacy issues and surveillance concerns through technology assessment and scenario planning. The Federal Trade Commission in the US requires companies to implement privacy protections in facial recognition from the design stage. Also, companies are supposed to iterate products to align with ethical standards based on anticipation.

‘Reflexivity’ recommends that the implementation of facial recognition technology in retail is monitored and evaluated. In 2020, IBM stopped providing facial recognition technology due to societal concerns regarding racial profiling. The governance practices and algorithms should be adapted based on reflections on the socio-ethical context of facial recognition and stakeholder feedback.

‘Inclusion’ emphasizes the importance of involving diverse stakeholders in the governance process, improving transparency and open innovation in each stage of facial recognition. Microsoft has been actively promoting transparency of facial recognition algorithms, and allowing third parties to review the systems, which enhances customer trust and regulatory compliance.

Finally, ‘Responsiveness’ involves being receptive to feedback from customers, organizations and society, such as Google’s Customer Feedback Systems and Ethical Committee. The feedback considers products and purposes, ensuring opportunities to meet customer demands and reduce social risks. Following these four dimensions goes a long way to foster the responsive development of facial recognition.

Figure 1: The roles of responsible innovation in governing facial recognition

Will facial recognition boost retail evolution?

If used properly, facial recognition will facilitate retail evolution, such as safer spaces, personalized shopping experiences and efficient operation systems. However, as a disruptive innovation in the early stage, facial recognition in retail and other sectors requires the involvement of governance strategies such as strict regulatory frameworks and responsible innovation.

About the author

Yalin Pang – A MSc Innovation Management and Entrepreneurship student at the University of Manchester in 2023-24. She is passionate about the intersection of emerging technologies, business model innovation, and sustainable development, and is keen on exploring the potential of cutting-edge technologies and business models to address global challenges.

An earlier version of this blog was prepared for BMAN73952 Global Challenges, Emerging Technologies and Governance, Alliance Manchester Business School, The University of Manchester.

Related