The contribution new AI rules: The EU’s billion-dollar dilemma first appeared at the online magazine Basic Thinking. You can start the day well every morning via our newsletter update.
Since August 2, 2025, new rules of the EU’s AI regulation-the so-called AI Act. Companies that offer AI models with “general use” must therefore fulfill certain information obligations. In our format “Break the News” we decrypted the background.
Background: AI Act
- The Ai act The European Union classifies AI models into different Risk level a. The following applies: The higher the risk of application for security, health or fundamental rights, the more strictly the requirements. Systems that represent a “clear threat” are completely prohibited.
- The regulation stipulates that for “AI models with a general intended use”, which will be released on the market from August 2, 2025, with which data the AI was trained. In addition, a strategy for dealing with copyrights must be recognizable. One applies to models that are already on the market Transitional phase Until August 2, 2027.
- In the event of violations of the AI Act, threaten from 2026 Fine of up to 35 million euros Or seven percent of the annual turnover. Such high punishments are initially questionable, because most major AI providers such as Google, Openaai and Anthropic have already volunteers “Code of Practice“Signed to the EU. Anyone who follows the code of conduct is considered innocent up to the proof of the opposite. Only Facebook group Meta refuses.
Classification: New AI rules of the EU
There are no limits to the possibilities of AI, unless politics does – and it undoubtedly has to. Because unregulated artificial intelligence A danger to democracy and society dar- to political manipulation, the violation of personality or data protection rights up to one complete loss of control via autonomous systems.
In its current form, however, the AI Act creates More uncertainty than securitybecause the EU planned guidelines not completed has a lot of room for interpretation. This in turn leads more bureaucracy – both on the EU and in many companies.
While it should be easy for the big AI providers to implement the new AI rules Start-ups and SMEs with great challenges. Because for them the current ambiguity represents a significant cost factor and additional effort they are cannot collect with billions in the hindquarters – Greetings from the GDPR.
Not the AI law itself, but the lack of implementation is not just one Innovation brakebut leads the basic idea, clarity, security and transparency to create an absurdity.
Voices
- Dirk Binding, expert in digital economy of the DIHK: “New documentation and evidence obligations also mean more effort and costs for entrepreneurs. That is why a strain on bureaucracy is immensely important.” At the moment, however, there are still no standards that have not been finished in time. This is concerned about uncertainty.
- Digital expert Anke Domscheit-Berg compare the AI Act with one Colossus on toned feet. Regarding a supervisory authority in Germany, she says: “The Federal Network Agency is expected to take on this task-this is consensus. So it will get another big task, without a lot of resources for this-and so it cannot use its supervisory function.
- Basic thinking editor-in-chief Fabian Peters Find: “The AI Act not only leaves a lot of interpretation, it is also holey in its current form. Therefore, it should be thought consistently – as a process that is at the end. Criticism is not only appropriate, but also desirable – regardless of which side. A fundamental defense against Meta is difficult to understand.”
Outlook: the EU’s billion dollar dilemma
The EU stands on a crossroads to take a credible one for AI Middle path between responsible regulation and innovation -friendly development to define.
Because: artificial intelligence can be properly regulated promote both social progress and economic strength – without pruning fundamental rights and security.
However, which currently does not fit together: The European Union wants Invest billions in AI – For example, in the form of gigantic data centers, which in turn fall under the AI Act.
Some of the planned investments that are intended to promote innovation would flow directly into the self -imposed rules. With it you cost yourself billions yourself. The AI regulation should therefore be continuously discussed and adaptable – just because of the rapid development of AI.
Also interesting:
- Controversial US software: Palantir monitoring in Germany?
- Chatgpt as a therapist: AI knows no confidentiality
- New format, exclusive content: Update newsletter becomes an independent product
- Irr Ki-Plan: Trump wants to rename artificial intelligence
The contribution new AI rules: The EU’s billion-dollar dilemma first appeared on Basic Thinking. Follow us too Google News and Flipboard Or subscribe to our update newsletter.
As a Tech Industry expert, I believe that the EU’s new AI rules present a significant challenge for the industry. While it is important to regulate AI to ensure ethical and responsible use, the strict regulations proposed by the EU could hinder innovation and stifle the growth of the industry.
On one hand, the rules aim to protect consumers and ensure that AI systems are transparent and accountable. However, the requirements for high-risk AI systems, such as facial recognition and biometric identification, could make it difficult for companies to comply and could ultimately drive up costs.
Additionally, the proposed fines for non-compliance are steep, with penalties reaching up to 6% of a company’s global revenue. This could have a significant impact on tech companies, particularly smaller startups who may struggle to afford such penalties.
Overall, while it is important to regulate AI to protect consumers and ensure responsible use, the EU must strike a balance between regulation and innovation. It will be crucial for policymakers to work closely with industry stakeholders to develop rules that are practical, enforceable, and promote innovation in the tech sector.
Credits