The AI Act – the artificial intelligence legislation – sets out the rules for the use of artificial intelligence

The AI Act (Artificial Intelligence Act) is a European Union regulation on artificial intelligence that came into force on 2 February 2025. The regulation governs the use of artificial intelligence within the EU. It came into effect on 2 August 2025. This world’s first AI Act was designed to set the rules for the development and application of AI technology.

Artificial intelligence has swept into all our lives at breakneck speed. Just a few years ago, we could not have even imagined the scale of this change, let alone its speed. This technology is so revolutionary that its possibilities are almost limitless, but at the same time it presents us with entirely new ethical and security issues. For these reasons, artificial intelligence needs its own legislation, which both promotes the responsible and transparent use of AI and protects us all from misuse.

We have entered the age of artificial intelligence, which requires new rules

This is not the first time that technological advances have revolutionized the way we operate. The first wave of digitalization brought us websites in the 1990s, the second brought search engines and social media in the mid-2000s, and the third brought the internet into our pockets through mobile technology in the 2010s. We have now entered the age of artificial intelligence, in which AI is no longer just one technological innovation among many, but is driving the next phase of digitalization.

Artificial intelligence is already everywhere. OpenAI released its first large language model in 2018, but the real breakthrough came in November 2022, when ChatGPT 3.5 became available. Since then, the role of artificial intelligence in everyday life and business has grown exponentially. This change is not only technological, but also societal and ethical, which is why artificial intelligence needs a set of rules. The AI Act is the EU’s response to this new era, and for the first time it establishes a globally influential framework for the development and use of artificial intelligence.

The content of the AI Act

The AI Regulation sets out requirements and obligations for developers and users of artificial intelligence. The rules relate to the intended uses of artificial intelligence and are designed to reduce the administrative and financial burden on businesses in relation to artificial intelligence.

Risk levels of artificial intelligence systems

The legislation classifies artificial intelligence systems into four risk levels, according to which requirements are imposed on them:

• Unacceptable risk: This includes systems that are considered a threat to people’s safety, livelihoods or rights.
  • The use of artificial intelligence for manipulation, fraud or the exploitation of vulnerabilities is not unequivocally permitted. Social scoring, predicting crime at an individual level and the unauthorised creation of facial recognition databases are also prohibited. Furthermore, artificial intelligence must not be used to recognise emotions in workplaces and educational institutions, for biometric classification or for real-time remote identification in public spaces for law enforcement purposes.
• High risk: This category includes use cases that may pose risks to health, safety and rights.
  • High-risk AI applications include, among others, security systems for critical infrastructure – such as transport and healthcare – as well as AI solutions used in educational institutions and the labor market that affect education and employment opportunities. Credit scoring, biometric identification and emotion analysis also fall into this category. Furthermore, the use of AI in law enforcement, migration management and the justice system is considered a high-risk area, as it can have a direct impact on fundamental rights and the rule of law. These systems must comply with strict requirements. Among other things, they must include risk management mechanisms and reliable data; users must be provided with clear information; and the system’s operation must be monitored under human supervision.
• Low risk (or risk relating to transparency): This category includes, for example, AI service bots, where the risk may relate to a person failing to recognise that they are interacting with artificial intelligence.
  • These systems are required to ensure transparency in the use of artificial intelligence in order to build trust. The AI Act requires that people be clearly informed when they are interacting with AI, so that they can make informed decisions. Certain AI-generated content, such as deepfakes and material intended for public communication, must also be clearly and visibly labelled.
• Minimal risk: Most AI systems fall into this category and are not subject to any requirements, such as spam filters and AI in video games.
  • The AI Regulation does not lay down rules governing the use of AI in these areas. However, openness and transparency in the use of AI are always encouraged.

How will the legislation affect general-purpose AI models (e.g. ChatGPT)?

General-purpose AI models, such as ChatGPT, are already integrated into many systems and services. They are powerful tools, but they can also pose risks. That is why the AI Regulation also sets out rules for the providers and users of these models to ensure that AI is safe and reliable to use. This includes, among other things, transparency requirements and copyright rules, and if a model poses significant risks, its provider must also identify and manage them. If your company uses ChatGPT or a similar AI model as part of its applications or services, it is important to ensure that its use complies with the regulations. For example, it must be made clear to users when AI is involved in decision-making, and copyright issues must be carefully taken into account.

How should we approach artificial intelligence in the future?

The AI Act brings order to the development and use of artificial intelligence. When a new AI system is launched on the market, responsibilities are clearly defined. The authorities ensure that AI complies with the rules, whilst users are responsible for ensuring that the system does not operate autonomously without human supervision, and providers of AI solutions must monitor the systems’ operation and address any issues promptly.

If artificial intelligence causes serious disruption or malfunctions, this must be reported without delay so that the adverse effects can be minimised. Furthermore, content generated by artificial intelligence must be identifiable; in particular, material intended for public communication must be clearly labelled.

When will the AI regulation come into full force?

The AI Act came into force on 1 August 2024, but its implementation will take place in stages. On 2 February 2025, the first part of the AI Act came into force, and AI systems posing an unacceptable risk, such as social scoring, are now officially banned in the EU. The rules and obligations for general-purpose AI models came into force on 2 August 2025. For high-risk systems, the transition period lasts until 2 August 2027.

The rules governing artificial intelligence are becoming clearer – and that’s a good thing

It is good that the ethics of artificial intelligence are being discussed, and that we are also establishing a common framework for all EU Member States to guide the ethical and safe use of artificial intelligence. The EU aims to be a global pioneer in safe artificial intelligence. By regulating AI in a way that respects human rights and fundamental values, the EU has the opportunity to create an AI ecosystem that benefits everyone. AI is developing rapidly, and it is important to prioritize transparency and accountability when utilizing it. Continuous monitoring and ethical practices ensure that AI operates reliably and brings genuine added value to businesses and society. Working with AI will therefore require greater transparency, accountability and continuous monitoring in the future, but at the same time it will create a safer and more reliable environment for the use of AI.

The ethical use of artificial intelligence is also a topic of discussion here at Hurja. That is why we have drawn up our own ethical principles for artificial intelligence, which guide its responsible use. AI can boost business efficiency, streamline processes and open up new opportunities, but only when used wisely. When utilized effectively, it supports better decision-making, boosts productivity and enables smarter digital services.

So, if you want to develop AI-powered solutions for your business in a smart and responsible way, let’s have a chat!