The EU AI Act for App Builders: What Changes for Your Product
Short answer: The EU AI Act regulates AI by risk level: a few practices are prohibited, high-risk systems face strict requirements, and many common app features such as chatbots or generated content mainly carry transparency duties. App builders should classify each AI feature, clarify whether they are a provider or deployer, train their team, and check current official guidance because timelines and guidance are still being refined.
On this page
Does the EU AI Act apply to your app?
If your app uses AI and is offered in the EU, or its outputs are used in the EU, the regulation likely applies in some form, regardless of where your company is based. The EU AI Act, Regulation (EU) 2024/1689, entered into force in August 2024 and applies in stages. It is not a ban on AI. It sorts AI systems by risk and attaches obligations to each level.
For most app builders the practical outcome is moderate: features like a support chatbot, AI writing help or image generation typically fall under transparency obligations rather than the heavy high-risk regime. The work is to know which features you have, which category each falls into and what role you play. This article is general orientation, not legal advice.
The four risk levels in plain language
Classification depends on the use case, not on the technology. The same language model can power a minimal-risk feature in one app and a high-risk system in another.
- Prohibited practices: for example manipulative techniques that cause significant harm, exploiting vulnerabilities of specific groups, social scoring, and certain uses of biometric identification and emotion recognition at work or in education. These prohibitions have applied since February 2025.
- High-risk systems: AI used in areas listed in the Act, such as recruiting and employee management, access to education, creditworthiness, access to essential public and private services, and some critical infrastructure, plus AI that is a safety component of regulated products.
- Transparency obligations: people must be told when they interact with an AI system such as a chatbot, synthetic content must be marked in a machine-readable way, and deepfakes must be disclosed, with some exceptions.
- Minimal risk: everything else, such as spam filters or recommendation features in many contexts, with no specific new obligations beyond general law.
Provider or deployer: which role do you have?
The Act distinguishes between providers, who develop an AI system or have it developed and place it on the market under their name, and deployers, who use an AI system in their own operations. An app company that builds an AI feature into its product and sells it is usually a provider of that system, even if it uses a third-party model underneath. A company that buys an AI tool for internal use is a deployer.
General-purpose AI model providers, such as the companies behind large language models, have their own obligations, which have applied since August 2025. If you build on their APIs, you rely on the documentation they provide, but you remain responsible for how your product uses the model.
Need a hand? neooutput helps with exactly this. Book a call for free and get a clear plan in 15 to 30 minutes.
Timeline: what is known and what to verify
Well-known milestones are the prohibitions and the AI literacy duty from February 2025 and the general-purpose AI model rules from August 2025. Most remaining provisions, including transparency obligations and many high-risk requirements, were scheduled for August 2026, with some product-related high-risk rules later. The European Commission has proposed adjustments to parts of the high-risk timeline, and guidelines, codes of practice and harmonized standards are still being published.
Because of this, do not plan against dates copied from an old blog post. Check the current text, Commission guidance and your national market surveillance authority before committing to a compliance plan, and revisit it regularly.
A practical checklist for app teams
- Inventory every AI feature in your apps and backends, including third-party SDKs that use AI.
- Classify each by use case: prohibited, high-risk, transparency or minimal, and document the reasoning.
- Clarify your role for each feature: provider, deployer or both.
- Implement transparency: label chatbots as AI, mark generated images, audio and video, and explain AI features in the interface.
- For anything potentially high-risk, start early with risk management, data governance, logging, human oversight and technical documentation.
- Train staff who build and operate AI features, which also covers the AI literacy obligation.
- Align with GDPR, since AI Act compliance does not replace data protection requirements.
Building AI literacy and governance
Article 4 requires providers and deployers to take measures to ensure a sufficient level of AI literacy among staff dealing with AI systems. For app teams, this means people understand how the AI features work, where they fail, which risks exist and which rules apply. neooutput offers workshops, trainings and 1:1 consulting on AI and agentic coding, remote or on-site in English and German, and can tailor sessions to product, engineering and compliance roles.
Organizations running many AI agents or AI features across departments also need governance tooling: an overview of systems, permissions, logs and approved models. kalAIDoCreate is an enterprise and public-sector AI orchestration platform for governed AI agents that supports this kind of oversight. A free call with neooutput is a good way to plan the training part.
Frequently asked questions
Is a chatbot in my app high-risk under the EU AI Act?
Usually not. A general support or product chatbot typically falls under transparency obligations: users must be informed that they are interacting with AI. It can become high-risk if it is used for decisions in listed areas, for example screening job applicants or deciding access to essential services.
Does the AI Act apply to companies outside the EU?
Yes, if they place AI systems on the EU market or if the output of their systems is used in the EU. This is similar to how the GDPR applies to non-EU companies. Non-EU providers of high-risk systems may also need an authorized representative in the EU.
What does AI literacy mean in practice?
It means people who build, operate or use AI systems on behalf of the organization understand their capabilities, limitations and risks well enough for their role. Typical measures are role-specific training, internal guidelines and documentation of who was trained. The required depth depends on context.
What are the penalties under the EU AI Act?
The Act sets fines that scale with the type of violation and company turnover, with the highest tier for prohibited practices. Lower tiers apply to other obligations, and there are adjustments for small and medium-sized companies. Check the regulation text and national enforcement guidance for details.
