Decisions that used to be made by people are now increasingly being made by AI.

States are also taking advantage of this.

How and to what extent will be depicted in this brief overview?

For this purpose, relevant terms are first explained before heading to two fields of

application, migration and policing, to discover how AI is used in governmental decision making.

Let's first head to the definitions.

Artificial intelligence is a term that is often used but still worth

defining as it may be hard to capture.

A possible definition is a system that is designed to operate with a certain level of autonomy and

that, based on machine and or human provided data and inputs, infers how to achieve a given

set of human-defined objectives using machine learning and or logic and knowledge-based approaches

and produces system-generated outputs such as content, predictions, recommendations or decisions,

influencing the environments with which the AI system interacts.

That definition is taken from the EU AI Act and European regulation on artificial intelligence.

Another important term is machine learning.

That is the implementation of computer algorithms that employ statistics to find patterns in data.

It is regarded as an application of AI and is often used for interference, classification,

prediction, forecasting and simulation.

To understand how those technologies can be used in government decision making,

let's first look into the field of migration management.

Around the world, governments increasingly try to regulate migration through automated systems.

Although it can be difficult to find out where and to what extent AI is implemented in those

systems and the decision making process, we want to look at some tested areas of implementation to

better illustrate what impact those technologies have on the individual.

AI systems can be implemented at every point of a migrant's journey.

At the first step, before the entry into the country of destination, machine learning systems

can be implemented to help streamlining visa application processes.

In most cases, they work with a traffic light system, assigning each applicant a green,

amber or red label. Green applications are supposedly low-risk applications

with a high chance of being accepted, while red applications are supposedly high-risk and likely

to be refused. A system like this was used by the British government up until 2020.

The factors which determined the risk scores for the individual applications were not disclosed,

but it can be assumed that nationality information was considered.

The system was trained on former application decisions and used its own decisions to reinforce

future classification. This resulted in a feedback loop until the system was declared

as racist and the government decided that it should be rebuilt.

At the next step, technologies used during entry are risk analysis to monitor borders like the

projects ROE Border or Nesta that were developed for the borders of the European Union. They are

meant to detect suspicious activity or persons in border regions.

ROE Border uses swarms of robots and drones equipped with cameras, while Nesta provides

border control officers with smart glasses that can alert them to areas with suspicious activity.

On top of that, there are technologies used to verify travelers' documents and narratives.

Among them are so-called automatic deception detection systems.

To understand those systems, it is important to know about biometrics first.

The GDPR, General Data Protection Regulation, defines biometrics as

personal data resulting from specific technical processing relating to the physical, physiological,

or behavioral characteristics of a natural person which allow or confirm the unique

identification of that natural person, such as facial images or dexillosophic data.

The latter is the investigation of the ridges of the inner surfaces of the hand and foot.

But now back to automation deception detection systems.