The biometrics and AI interface

Unprecedented disagreement has emerged within the global biometrics community regarding the intricate relationship between biometrics and artificial intelligence (AI). A new paper from the Biometrics Institute, “Members’ Viewpoints: The Relationship Between Biometrics and Artificial Intelligence (AI)” explores the conflicting perspectives gathered throughout 2024 from members and other experts across the globe. This highlights the complex connections between these two rapidly evolving technologies, revealing a landscape of contrasting views within the field.

Opinion between AI and biometrics is divided

The paper highlights significant disparities in member opinions regarding the relationship between biometrics and AI.

Some members argue that biometrics are inherently intertwined with AI, while others emphasise that many biometric applications exist independently.

“Rarely has the biometrics community disagreed on an issue at this level before,” says Isabelle Moeller, CEO of the Biometrics Institute, “This paper reflects the conflicting perspectives of our global community on an evolving topic that is critical technology for biometric success. Understanding the relationship between biometrics and AI is essential for responsible innovation and the development of ethical guidelines for their use.

While ISO definitions for both biometrics and AI exist, other important but sometimes non-aligned definitions are prevalent in the public domain. A diverse range of stakeholders, including major technology corporations, civil society organisations, also seek to define AI. The Biometrics Institute offers its own perspective to address this, and the paper resulted in a new entry for “Artificial Intelligence” in the Institute’s Explanatory Dictionary of Biometrics, providing a resource that reflects these various viewpoints for the benefit of members, policymakers, and the general public.

Bridging the gap: Defining biometrics and AI

Defining biometrics and artificial intelligence (AI) presents significant challenges. Existing definitions, such as those from ISO and some governments, are for some audiences too technical, complex, or inaccessible, affecting broader understanding. The Biometrics Institute’s Explanatory Dictionary aims to bridge this gap by capturing the nuances of these terms, considering both formal definitions and how they are perceived by the public – for example, how someone might explain biometrics or AI to a friend. However, there are no universal definitions of biometrics or AI and public perception of these technologies remains unclear, influenced by often-confusing social and traditional media portrayals that frequently fail to distinguish between different types of biometrics and AI, creating a sense of technological ambiguity.

AI’s impact: Enhancing or endangering biometrics

Determining which biometric applications “have” AI can be challenging, as there are many ways that AI interacts with biometric technologies, from aiding in processing as a threat vector to enhancing security measures and general processes. However, AI also presents potential risks, such as vulnerabilities that could be exploited. As AI becomes increasingly integrated into all aspects of technology, identifying and assessing these risks will become more crucial. Ultimately, the increasing complexity of AI integration may make it difficult to distinguish AI components from other elements within a given system.

The relationship between biometrics and AI: Complex and multifaceted

While some argue that biometrics are inherently intertwined with AI, others emphasise their potential for independent use. The inclusion of AI in a specific biometric system is often determined by the application’s requirements, rather than the type of biometric data used. Furthermore, the role of human operators within biometric systems highlights the ongoing debate surrounding human error versus machine error and the varying levels of acceptance for each. This debate is further complicated by differing legislative and liability frameworks.

Regulation and innovation: Challenges will persist

Existing regulations often conflate these two distinct technologies, particularly in the case of face recognition, where the term “AI” is frequently used interchangeably. This ambiguity can lead to overly broad regulations that may inadvertently restrict the development and deployment of beneficial biometric applications. While concerns regarding the potential impact of remote biometric surveillance on civil liberties are valid, some argue that excessive regulation could stifle innovation and hinder the ability to address pressing societal challenges through the use of cutting-edge technological solutions.

The paper’s key findings provide many more insights into the definitions of biometrics & AI, the impact and influence of AI on biometric processes, the relationship between biometrics and AI and regulatory oversight of biometrics and AI.

See Addendum below

For a deeper dive into the broader implications of AI on biometrics, readers are encouraged to consult the Institute’s State of Biometrics Report. The report explores key themes including “To AI or not to AI?” And “Is AI outsmarting us?” With a strong focus on responsible use, and the need to put people first to build public trust in biometrics, the report also explores critical considerations such as the importance of privacy and security and offers a comprehensive perspective on the evolving landscape of these technologies.

The full paper, “Members’ Viewpoints: The Relationship between Biometrics and Artificial Intelligence (AI),” is publicly available. The Institute will continue to discuss this crucial relationship at upcoming events, including the Asia-Pacific Conference in Sydney in May, its US Strategy Forum, and The Impact of AI on Biometric Vulnerabilities Workshop in New York in June.

ENDS

Addendum – Headline Takeaways

Definitions of biometrics & AI

• There are no universal definitions of biometrics or AI and those put forward by ISO and some governments are either too technical, obtuse or are not fully aligned with one another or are hidden behind paywalls and not accessible to the majority of the general public
• The role of the Biometrics Institute’s Dictionary is in capturing the shades of meaning and the different perspectives of key terms such as biometrics and AI – the definitive meaning and the general perceptions e.g. explaining the terms to a neighbour at a BBQ
• Therefore, what is the public’s perception of biometrics and AI? Good, bad or ugly?
• The media (social and traditional) struggle to define, explain and differentiate between biometrics, in all its forms, and AI, in all its forms. A technological soup…

Impact and influence of AI on biometric processes

• Which biometric applications ‘have’ AI? What do we mean by that?
• There are many ways that AI interacts with biometric technologies, to aid processing, as a threat vector, as a protective measure or as an enhancement to general processes. Therefore, do they all present a serious risk, or just some?
• As AI becomes more pervasive in all technologies how will we identify and assess the various risks? Eventually, will we even be able to separate ‘AI’ components from any other element in a system?

 Relationship between biometrics and AI

• There are conflicting opinions among members regarding the relationship between AI and biometrics
• Some say that biometrics are an adjunct to AI technology and as a consequence are always an integral part of it
• Others point out that while biometrics and AI can be used together, in a variety of applications, it is also the case that some biometric applications exist quite separately from AI
• The inclusion of AI in any biometric process is dependent on the use case and not necessarily the biometric modality. Systems using the same biometric modality may or may not employ AI technology subject to the operating requirements
• Human operators working within biometric systems have also been cited as examples of non-AI processing but their continued existence and contribution ​​​​depends heavily on their ability to exceed or even match the future performance of AI software.  There is a heavy dependence on legislative and liability treatment – the criteria appear to be different. There seem to be one set of views around humans making mistakes versus another set that are overly critical of mistakes made by machines.

Regulatory oversight of biometrics and AI

• The terms biometrics (especially the use of live and remote biometric surveillance) and AI have become conflated in some regulation e.g. the EU AI Act
• ‘Face recognition’ and ‘AI’ ​​​​are virtually interchangeable in some contexts. The blanket term “AI” is frequently used when “face recognition” would be the correct term. This might also be applied to other biometrics, especially speaker authentication
• Some regard all remote biometric techniques as restrictive of civil liberties and a threat to basic human rights
• Others consider some aspects of regulation to be excessive because it potentially constrains innovation, entrepreneurship and the ability to deploy cutting edge technological solutions to current societal problems and challenges

About the Biometrics Institute 

The Biometrics Institute is the independent and impartial international membership organisation for biometric users and other interested parties. It was established in 2001 to promote the responsible, ethical and effective use of biometrics and has offices in London and Sydney. 

The member register which represents a global and diverse multi-stakeholder community now lists over 200 membership organisations from 41 countries. It includes banks, airlines, government agencies, biometric experts, privacy experts, suppliers, academics and 18 Observers representing United Nations agencies, IGOs and European Union institutions. 

The Biometrics Institute connects the global biometrics community. It shares knowledge with its members and key stakeholders and most importantly, develops good practices and thought leadership for the responsible, ethical and effective use of biometrics. 

For more information, please email Marco Lombardi.