AI Development Services

EU AI Act: What It Means for Your App Project

For most business apps with AI features, the EU AI Act mainly means transparency duties, such as telling users they are talking to an AI and marking AI-generated content, plus AI literacy for the people who build and run the system. Stricter obligations apply only if your use case falls into a high-risk category, such as hiring, credit scoring, education or critical infrastructure, so the first job is to classify your project correctly.

neooutput editorial team 4 min read Auf Deutsch lesen

The risk-based structure in one paragraph

The AI Act sorts AI systems by risk: prohibited practices, high-risk systems, systems with specific transparency obligations, and minimal-risk systems with no specific new duties.

Prohibited practices include, for example, manipulative techniques that cause significant harm, social scoring and certain uses of biometric identification and emotion recognition in workplaces and schools. High-risk systems are listed in the regulation's annexes and include AI used in recruitment, worker management, access to education, creditworthiness assessment, essential public services and safety components of regulated products. Separately, providers of general-purpose AI models, such as the large language model vendors, carry their own obligations. Most app projects use those models rather than provide them.

Provider or deployer: know your role

Your obligations depend on whether you are a provider (you develop an AI system and place it on the market under your name) or a deployer (you use an AI system in your business).

A company that builds an AI feature into its own product for customers is typically a provider of that AI system, even if it uses a third-party model underneath. A company that uses an AI tool internally is typically a deployer. Your development agency usually builds on your behalf, so the role normally stays with you. Clarify this in the project contract.

Transparency duties that affect many apps

The obligations most app projects run into are the transparency rules, which apply according to the regulation's staged timeline.

Users must be informed when they interact with an AI system, unless it is obvious from context. Synthetic audio, image, video and text content generated by AI systems must be marked in a machine-readable way where technically feasible. Deployers of deepfakes and of AI-generated text published to inform the public on matters of public interest have disclosure duties, with some exceptions. In practice that means clear labels in chat interfaces, metadata or watermarks for generated media, and a short explanation in your app.

  • Label chatbots and AI assistants clearly.
  • Mark generated images, audio and video, for example via metadata.
  • Explain AI features in your app and privacy notice.
  • Keep records of which models and versions you use.

If your use case is high-risk

High-risk classification brings substantial obligations, so check it early and with legal support.

Providers of high-risk systems need a risk management system, data governance for training and test data, technical documentation, logging, transparency information for deployers, human oversight measures, accuracy and robustness requirements, a quality management system and a conformity assessment before placing the system on the market. Note that the timeline for high-risk obligations has been subject to political discussion about postponement, so check the current status for your project rather than relying on older articles.

AI literacy and good practice for everyone

Since February 2025, providers and deployers must take measures to ensure sufficient AI literacy among staff dealing with AI systems.

That is a good reason to train your team, not only on using AI tools but on their limits and risks. Beyond formal duties, practices that the AI Act encourages are simply good engineering: document which models you use and why, evaluate outputs systematically, log decisions, keep a human in the loop for consequential actions, and have a process for reporting and fixing problems. Agentic coding workshops, such as those neooutput offers, can contribute to building this literacy in development teams.

A practical AI Act checklist for app projects

Run through these points at project kick-off and again before launch.

  1. Describe each AI feature and its purpose in one sentence.
  2. Check it against the prohibited practices and the high-risk annexes.
  3. Determine your role: provider, deployer or both.
  4. Plan transparency measures in UI and content.
  5. Document models, data sources, evaluation and human oversight.
  6. Align with GDPR documentation, since the two overlap heavily.
  7. Assign an owner for monitoring regulatory updates.

Frequently asked questions

Does the AI Act apply to companies outside the EU?

Yes, if they place AI systems on the EU market or the output of their AI systems is used in the EU. Like the GDPR, it has extraterritorial reach.

Is a customer service chatbot high-risk?

Usually not. A typical support chatbot falls under the transparency obligations, meaning users must know they are interacting with AI. Context matters, so check special cases.

Do I need a lawyer for AI Act compliance?

For high-risk or borderline cases, yes. For typical apps with transparency duties, a good development partner can implement the measures, while legal review confirms the classification. This article is not legal advice.

All insights