ICT is becoming simpler and more human also when it comes to artificial intelligence (AI). With the advent of the European AI Act organizations in Europe must be more careful with AI applications. The law is being introduced in stages and since Aug. 2, 2025 new obligations apply. What has changed and what does it mean for your organization?
What did the first phase entail?
In February 2025, the first phase of the AI Act was launched. The focus then was mainly on:
- Awareness and basic knowledge of AI in organizations.
- Avoiding prohibited AI applications (the category unacceptable risk).
- Underlining that not only developers are responsible, but also providers, users and other links in the chain.
What has changed since August 2025?
Since August, a new step has been taken. In particular, organizations working with generative AI (such as chatbots or AI assistants) will have to deal with this.
Key commitments:
- Transparency Users need to know that they are interacting with AI, including clear notification that AI can make mistakes.
- Human supervision AI should not just make decisions that have major consequences. An employee should always be able to intervene and assess.
- Documentation & accountability Organizations must document how models work, what data they are trained with and how risks such as bias are mitigated.
What does this mean for different roles?
| Role | Obligations |
|---|---|
| Developer/provider | Ensure transparency, security and technical requirements. |
| User | Deploy AI responsibly, inform employees and test outcomes. |
| Importer/Representative | Ensure AI products comply with the law and enable audits. |
Many companies perform multiple roles simultaneously - and with them, multiple responsibilities.
Outlook: 2026
From August 2026 is when things get really exciting. Then the rules for high-risk AI applications, for example, in healthcare, education, justice or security. This category has strict monitoring, testing and documentation requirements.
Those who start now with good processes and clear agreements will soon be a lot stronger.
What can you do now?
- Map what AI applications exist within your organization.
- Check which risk category they fall under.
- Make sure employees know when they are dealing with AI.
- Always maintain human supervision when making important decisions.
- Document the use of AI and continue to monitor.
Our vision
At Analyst ICT, we see AI as an opportunity, if applied properly. The AI Act helps to create clarity and handle AI responsibly. By taking the right steps now, you avoid stress and hassle the moment the stricter rules take effect. Do you enjoy learning about AI? Or do you need help with the proper application in this? Please get in touch.
