Innovations in technology have brought about significant changes in police lineup procedures, with the potential to improve both the accuracy and fairness of criminal lineup identifications. However, along with these advancements come ethical challenges that require thoughtful consideration. Value Sensitive Design (VSD) is a design approach that prioritizes the incorporation of human values into the evolution of technology. In what follows, I explore the application of VSD to lineup technologies.
Value Sensitive Design and Moral Imagination: A Primer
Value Sensitive Design is an approach that encourages the inclusion of human values in the design and development of technology. Central to VSD is the concept of moral imagination, which urges designers to think carefully about the potential ethical implications of their creations and proactively address them. Moral imagination helps designers—which to my mind, includes people like me who research eyewitness identification procedures and applications that facilitate remote victim accounts of GBV—to identify possible consequences and develop strategies to mitigate negative impacts on individuals and society.
Of course, human values are complex, multi-dimensional, and subject to interpretation. They are personal, as well as cultural and societal, and contested; hence human values change over time. Therefore, it is crucial that we bring together different stakeholders to understand the diverse and often conflicting human values in technology design.
Applying Moral Imagination to Person Identification Technology
When it comes to police lineup technology, the concept of moral imagination urges us to strike a balance between designing procedures that maximize accuracy and fairness while at the same time maintaining ethical principles. By considering potential ethical issues in the development of new lineup technologies, designers can create solutions that not only improve the outcomes of the process but also better protect the rights and dignity of those involved.
The challenge of incorporating moral imagination into lineup technology can be seen in the development of digital camera and facial recognition algorithms. Some of these algorithms struggle to accurately capture and identify individuals with darker skin tones, increasing the risk of wrongful identification. To illustrate, following George Floyd's murder in Minneapolis in 2020, several cities and states banned or limited law enforcement's use of facial recognition. Yet, places like Detroit continued using it despite growing concerns.
Consequently, a Black man named Robert Williams was wrongfully arrested in Detroit after a facial recognition software error falsely matched his driver's license photo to a shoplifting suspect. This incident shows the risks of over relying on technology and the urgent need for unbiased, accurate systems. As designers, we must confront this issue head-on and work diligently to create algorithms that equitably and accurately represent people of all ethnicities.
As another example, own race bias, which refers to our ability to recognize individuals more accurately if they are from our own race/ethnicity, is one of the largest contributors to lineup identification errors. The leading theories on this phenomenon suggest that the difficulty arises from a lack of social contact and perceptual expertise, resulting in greater challenges in learning and remembering the facial features of other race faces. I wonder whether police digital technology might also play a role. Specifically, are the digital camera algorithms capturing Black suspects' facial features in enough detail for reliable eyewitness identification? This unaddressed question requires research and thoughtful action to correct any deficiencies.
As designers, we must have the moral imagination to acknowledge these biases and actively work to eliminate them from algorithms. Collaboration with diverse communities and experts can help to ensure the technology is more equitable and less likely to perpetuate systemic injustice.
Privacy and consent are also crucial considerations in lineup technologies that involve the collection and storage of personal data. For instance, imagine a man who was convicted of petty theft a decade ago and has maintained a clean record since then. Suppose his mugshot is used as a filler in a photo lineup for a murder case without his consent or knowledge. This breach of privacy endangers his reputation, and it places him at risk of being falsely accused. We must have the moral imagination as designers to take steps to safeguard individuals' data to protect individuals and maintain the public's trust in the systems we develop.
Considering the Ethical Implications of Lineup Technology Innovations
The ethical implications of new lineup technologies extend beyond the design stage. Stakeholders, including law enforcement agencies, legal professionals, and policymakers, must also think about the ethical consequences of adopting and implementing these innovations.
As an example, we have been extensively investigating interactive lineups, which allow witnesses to explore the lineup members from multiple angles. However, this advance might unintentionally introduce new biases or distractions that could affect witness decisions and how police conduct the procedure. As such, we must these new presentation methods rigorously to identify and address any unintended consequences and ensure a rigorous and fair identification process.
Relatedly, our lab is collaborating with Dr Michael Gaebler from the Max Planck Institute for Human Cognitive and Brain Sciences, Professor Peter Eisert from the Humboldt- Universität zu Berlin, and Dr Anna Hillsmann from Fraunhofer Heinrich Hertz Institute, Dr Anna Bobak from the University of Stirling, and Dr Steve Lindsay and Eric Mah from the University of Victoria in an ESRC/DFG/SSRH funded project that explores the use of virtual reality (VR) technology in interactive lineups, which introduces new ethical concerns. While VR technology can offer a more immersive experience for witnesses, potentially helping them to better recall and identify suspects, it also raises questions about the potential for psychological distress. As VR technology becomes increasingly realistic and immersive, the possibility of retraumatizing witnesses must be carefully considered and addressed.
Ensuring high fidelity in a virtual reality identification system is also paramount. The facial features and likeness of individuals in the VR environment must accurately reflect real-life appearances to help establish a reliable link between the police suspect and the identification outcome.
The potential for misuse is another ethical issue that arises from the adoption of new lineup technology. AI-generated deep fake faces as fillers in police lineups could lead to overreliance on technology and compromise human discernment. Biased lineups results from face generating systems that are built on biased face datasets, and therefore, we must not have blind faith in these systems. It's essential for researchers, law enforcement agencies and legal experts to stay cognizant of these technologies' limitations and guarantee that human judgment isn't wholly supplanted.
As a final example, during the Covid-19 lockdowns in the UK, we discussed extensively with our police partners the potential for conducting lineups remotely using video conferencing technology. The pandemic brought about significant case backlogs, and remote lineups potentially could save time and resources for law enforcement and witnesses. However, remote lineups raise concerns about the integrity of the process and the potential for outside influence on witness decisions that could tamper with the identification process. (For wider reading, I recommend this new report on Blended Advice and Access to Justice and this one on remote Crowne Court hearings.)
Conclusion
Value Sensitive Design and moral imagination offer a valuable framework for addressing the ethical complexities of developing and implementing new lineup technologies. By having the moral imagination to consider the potential ethical implications from the design stage, and throughout the adoption and use of these technologies, stakeholders can collaborate to create innovative solutions that uphold ethical principles and promote justice. Embracing VSD and moral imagination ensures that as technology continues to advance, so does our commitment to fairness and the protection of individual rights.
For further reading:
Value Sensitive Design: Shaping Technologies with Moral Imagination by Batya Friedman and David G. Hendry
Anatomy of an AI System by Kate Crawford and Vladan Joler
Comments