UAE Public Prosecution’s 2045 Vision keeps humans in command of AI

Abu Dhabi: “Human will be in command each and every time.”

For Counsellor Salem Al Zaabi of the UAE Public Prosecution, this principle defines how artificial intelligence should be introduced into criminal justice: not as a substitute for human judgment, but as a capability operating within clear legal, ethical and institutional boundaries.

Speaking during the 15th United Nations Congress on Crime Prevention and Criminal Justice in Abu Dhabi, Al Zaabi outlined the UAE Public Prosecution’s approach to the future of criminal justice, and the role artificial intelligence can play in strengthening institutional capability while preserving human responsibility and the rule of law.

At the heart of that approach is a clear principle:

Technology may enhance the capabilities of justice institutions, but responsibility for judicial and prosecutorial decisions remains human.

A system transformation, not a technology upgrade

Al Zaabi described the UAE Public Prosecution’s Future Prosecution 2045 vision as a system-wide transformation rather than the introduction of individual technological tools.

The vision examines how digital technologies, artificial intelligence and other emerging capabilities can strengthen the effectiveness, accessibility and future readiness of criminal justice while preserving its constitutional, legal and institutional foundations.

In this model, AI is an enabling capability within the justice system — not the system itself.

The objective is therefore not simply to introduce more technology, but to develop a justice system capable of using technology responsibly, while remaining firmly anchored in the rule of law.

Trust at the centre of transformation

For Al Zaabi, trust is fundamental to the relationship between justice institutions and society.

As AI becomes increasingly integrated into justice processes, maintaining that trust requires more than technical performance. It requires clear ethical principles, institutional policies, accountability mechanisms and effective human oversight.

This is why the UAE Public Prosecution’s approach brings together strategy, ethics, policy, governance and oversight within an integrated institutional framework.

The purpose is to ensure that technological capability develops alongside — and never ahead of — the safeguards required to govern it.

Where AI stops and human authority begins

A central feature of the Public Prosecution’s model is the clear distinction between tasks that technology may perform autonomously, tasks in which AI may provide analysis or recommendations, and decisions that must remain exclusively human.

AI can support prosecutors by retrieving information, organising case material, analysing data, identifying patterns, generating summaries and providing structured recommendations.

But final determinations that carry significant legal consequences remain subject to human authority, professional judgment and institutional accountability.

This distinction goes beyond simply placing a person “in the loop.”

The Public Prosecution’s model provides for different levels of human control according to the potential impact and risk of the AI-supported task, with stronger safeguards and greater human intervention as the potential effect on rights and legal interests increases.

As Al Zaabi put it:

“Human will be in the loop, on the loop, and in command.”

The underlying principle is clear: AI may support judgment. It does not assume the non-delegable responsibility attached to judicial decision-making.

AI as a new stage in the evolution of criminal justice

During the session, Al Zaabi placed artificial intelligence within the wider historical evolution of criminal justice.

Previous transformations reshaped fundamental elements of justice — from the transition from private retaliation to public justice, to the development of fair-trial guarantees and the introduction of scientific evidence.

Artificial intelligence represents a particularly significant transformation because its impact potentially extends across multiple dimensions of the criminal justice system at the same time — including crime, evidence, investigation, decision-making and procedural safeguards.

The lesson from previous transformations, Al Zaabi explained, is not that change should be rejected, but that it must be properly governed.

The UAE Public Prosecution’s approach therefore begins with a different question.

Not simply: What can artificial intelligence do?

But: What should artificial intelligence be permitted to do, what role should it perform, and what must remain the responsibility of humans?

Governance first

One of the defining features of the UAE Public Prosecution’s experience is the sequence in which its approach to AI has been developed.

The institution did not begin with technology and then seek to construct safeguards around it.

It sought first to establish the strategic, ethical and institutional framework within which technology could operate.

This includes an AI strategy, ethical principles, institutional policies, oversight mechanisms, risk-based controls and capability-building.

The principle can be expressed simply: Governance first. Technology deployed within governance.

Asked how public confidence in artificial intelligence in justice can be strengthened, Al Zaabi’s answer was direct:

“Governance, governance, governance.”

For the Public Prosecution, governance is not an additional layer applied to technology after deployment. It is part of the architecture through which AI is designed, approved, operated, monitored and reviewed.

From governance to practical application

The session also demonstrated how this approach is moving from institutional frameworks into practical applications.

Specialised AI agents can support prosecutors in functions such as case management, information retrieval and file summarisation through controlled access to authorised information and systems.

Rather than requiring prosecutors to manually navigate multiple records for routine information, AI-supported tools can help retrieve, organise and present relevant information in a more efficient and structured manner.

The objective is not to transfer legal judgment to machines.

It is to reduce the burden of information-intensive and routine tasks, allowing prosecutors to dedicate greater attention to the aspects of their work that require legal reasoning, professional assessment and human judgment.

The operational principle remains: AI retrieves, organises, analyses and proposes. The human reviews, decides and remains accountable.

Risk-based governance

The UAE Public Prosecution’s approach also recognises that not every use of artificial intelligence carries the same level of risk.

Administrative and routine applications require a different level of control from tools that may support analysis related to investigations or other matters with potential legal consequences.

Accordingly, the model applies graduated levels of governance and human intervention depending on the impact of each AI-supported function.

The greater the potential impact on rights, liberty or legal interests, the stronger the required safeguards and the higher the level of human control.

This risk-based approach is intended to ensure that technological capability develops in proportion to legal responsibility.

A human-led, AI-enabled justice system

The UAE Public Prosecution’s vision is therefore not one of autonomous justice. It is a vision of a human-led justice system strengthened by advanced technological capabilities. Human judgment remains central. Legal authority remains defined by law. Institutional responsibility remains traceable. And technology remains subject to governance.

This distinction is critical to the Public Prosecution’s wider vision for 2045: technology may transform the way justice institutions operate, but it does not alter the constitutional and legal foundations from which their authority derives.

From national experience to international dialogue

Al Zaabi also linked the UAE Public Prosecution’s experience to the wider purpose of the United Nations Congress.

As patterns of crime evolve and increasingly involve digital technologies, complex networks and cross-border dimensions, justice institutions face challenges that cannot be addressed in isolation.

International dialogue provides an opportunity to exchange experience, understand different approaches and develop stronger cooperation between justice institutions.

As Al Zaabi noted, changes in the nature of crime make it increasingly important for institutions to talk to each other, listen to each other and learn from each other.

For the UAE Public Prosecution, participation in the Congress is therefore not limited to presenting its own experience.

It is an opportunity to contribute to a wider international conversation on how criminal justice systems can adapt to technological transformation while preserving the rule of law, legal safeguards and human accountability.

The measure of success

Ultimately, the success of artificial intelligence in justice will not be measured by how extensively technology is deployed.

It will be measured by whether technology improves the ability of justice institutions to fulfil their responsibilities while preserving public trust, legal safeguards and human accountability.

For the UAE Public Prosecution, the future of AI in criminal justice can therefore be captured in three principles: Human in the loop. Human on the loop. Human in command.

And behind those principles lies a broader institutional commitment:

Justice remains the governing reference for technology — not the other way around.