AI Act’s Impact on Law Enforcement: Biometric Expert’s View

The AI Act, proposed by the European Commission in April 2021 and passed in March 2024, aims to establish a comprehensive framework for AI systems within the EU. It categorizes them based on their potential risk levels, ranging from minimal to unacceptable, and imposes regulatory requirements accordingly. Notably, the Act emphasizes transparency, accountability, and human oversight in AI deployment to safeguard fundamental rights and societal values.

We spoke with Vincent Bouatou, Deputy Chief Technology Officer at IDEMIA Public Security (IPS), to find out more about the impact of the AI Act and how it will influence the future of innovation.

Can you explain more about the AI Act and its potential repercussions for the technology industry?

At IPS, we strongly support regulatory initiatives directed toward ensuring public trust and the responsible use of technologies and have been actively engaged in discussions surrounding these topics. The Act provides much-needed clarity and regulation for technology providers, customers, and end users, particularly concerning use cases and data governance practices.

Additionally, the AI Act introduces stringent checks and audits, mandating thorough documentation of product deployments and algorithmic developments. This encompasses data quality assurance and risk management measures across both development and deployment phases. These requirements highlight the need for meticulous scrutiny of processes previously left to self-regulation or open to interpretation.

Compelling technology providers to explain how their systems operate creates a more understandable product ecosystem. This would enable organizations using the systems to address any concerns end users and data subjects may have, ultimately fostering trust and reinforcing responsibility.

Moreover, we believe the AI Act will benefit the technology industry by promoting unbiasedness and fairness in AI systems. While there may be challenges in adjusting to the new regulations, compliance is essential in maintaining the industry’s reputation and securing sustainable growth.

Who will guarantee that each EU Member State respects the AI Act regulations?

The Act includes the creation of national market surveillance authorities tasked with overseeing adherence within their respective jurisdictions. Their role encompasses monitoring implementation, conducting inspections, and enforcing compliance. Furthermore, they will engage in collaborative efforts with other national and EU-level agencies to maintain consistency in enforcement and interpretation across Member States.

In our capacity as an industry leader, we are committed to actively participating in discussions to facilitate understanding of the industrial state-of-the-art and technological advancements among relevant authorities. This commitment is exemplified by our past collaborations with entities such as the French Commission Nationale de l’Informatique et des Libertés (National Commission on Informatics and Liberty) and the National Institute of Standards and Technology (NIST) in the USA on matters specific to biometric technologies.

Additionally, regular exchanges among all stakeholders are being conducted. They are essential to stay up to date and potentially refine regulation, for instance, through delegated acts. These ongoing dialogues help maintain the regulatory framework’s responsiveness to emerging challenges and technological advancements.

What steps is IPS taking to ensure all its solutions align with the requirements outlined in the AI Act?

We are looking at the gaps in our processes regarding the AI Act and are very pleased to find few. Apart from some minor work on documentation of our product and development practices, our solutions and internal processes are already, in essence, aligned with almost everything stipulated in the Act.

IPS is headquartered in France. We have one of the biggest R&D teams in a corporate setting—also located in France—contrary to many of our competitors. All our solutions and development processes already comply with GDPR.

We strive to produce algorithms that are fair and unbiased to everyone, regardless of their age, ethnicity, or gender. However, this is not a straightforward process and, to accomplish this, our algorithms are exposed to rigorous testing. We are proud to say that our biometric identification algorithms have reached such a high level of efficiency that biases can no longer be measured—there are not many biometric companies that can make such a bold statement and back the claim up with independent testing.

For many years, IPS has participated in NIST’s benchmark testing for three biometrics: iris, face, and fingerprints. We have always ranked either first or in the top tier of these benchmarks—a great feat for any technology provider.

How will the AI Act impact law enforcement agencies in the EU?

One of the most significant impacts lies in the Act’s emphasis on accountability and transparency. Law enforcement agencies deploying biometric systems must provide clear explanations of how they procure and operate their biometric identification systems, including algorithms, data governance, and potential biases.

The regulations in the AI Act also provide clarity on the permissible use of video analytics platforms for law enforcement, outlining precise guidelines for their utilization, with a short and clear list of exceptions to the general prohibition of real-time biometric identification in public spaces.

By selecting IPS as their partner of choice, law enforcement agencies gain assurance that they are teaming up with a European technology provider that is fully compliant with GDPR, and the regulations outlined in the AI Act. We have been fully engaged in the development of the regulations as the voice of the industry, setting us apart from competitors.

We bring extensive experience to the table, having championed this cause for years. Leveraging our technology instills confidence and peace of mind for law enforcement agencies worldwide. Our algorithms, developed in Europe, stand out as the fairest and most accurate on the market, which is a difficult and rare combination that also sets us apart from the competition.

What about law enforcement agencies outside of Europe?

For law enforcement agencies beyond the EU, collaborating with IDEMIA Public Security sets a benchmark, as they adopt compliant technology ahead of potential regulatory shifts. This proactive approach positions them ahead of the curve, in a climate where other regions may follow suit with their own versions of the AI Act.