Is your startup developing or testing an AI-based solution? Do you have a working MVP, acquired first users and thinking about scaling? Before you invest in further development or enter discussions with VCs, you need to know one thing: Your system may be classified as “high risk” under the AI Act. And that means specific obligations and potential sanctions.
AI Act and MVP? Yes, it is combined
Most founders think that regulations will only start affecting them “someday.” Meanwhile, the AI Act classifies systems as early as the design and testing stages. An MVP with AI that recommends products, analyzes resumes, supports credit decisions or manages user behavior could end up in the high-risk systems category.
How do you know if your AI is a “high-risk system”?
The AI Act indicates the areas in which working AI systems are regulated most heavily. Examples:
- biometric data processing,
- credit and employment scoring,
- Managing the behavior of students or employees,
- public space video analysis,
- Product recommendations that have a significant impact on user decisions.
If your MVP:
- analyzes personal data,
- makes decisions that affect user rights,
- Personalizes content or offers,
- …There is a high probability that it falls under the AI Act.
What are the penalties for non-compliance with the AI Act?
Penalty of up to €30 million or 6% of annual turnover (whichever is greater),
Withholding the system from the market,
Requiring documentation, audit, risk assessment, list of data providers, explanation to users.
How do you check MVP compliance with the AI Act?
- Identify the uses of AI in your MVP – what decisions does it support?
- Evaluate whether data is processed automatically and whether it affects people.
- Check whether your use case is in the high-risk catalog (e.g., according to the Annex III AI Act).
- Assess potential risks (e.g., discrimination, lack of transparency, manipulation).
- Initiate AI management documentation and policies in accordance with ISO 42001 or TR 38507.
What is worth doing before scaling?
- Have an independent audit of AI Act + RODO compliance,
- Prepare a compliance checklist for VC interviews,
- Develop an AI Policy that shows that your company is thinking about responsible technology development,
- If necessary – adjust the MVP before you overinvest.
Compliance is not a brake, but the foundation of a scalable business
Implementing AI in a startup doesn’t have to mean red tape. But lack of awareness of regulations is a real risk to investment, reputation and growth.
Act ahead – make sure your MVP is AI Act compliant now, before repair costs exceed construction costs.
Sources:
https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai
https://kpmg.com/ch/en/insights/artificial-intelligence/iso-iec-42001.html
https://artificialintelligenceact.eu/
Leave A Comment