2 August 2026 marks another major milestone in the implementation of the European Union’s AI Act. New obligations are now applicable for providers of general-purpose AI (GPAI) models, while transparency requirements for AI-generated content also become enforceable across the EU. Together, these measures represent an important step towards making AI systems more transparent, accountable, and worthy of public trust.
The AI Act is the world’s first comprehensive legal framework regulating artificial intelligence through a risk-based approach. Rather than regulating all AI equally, the legislation introduces obligations proportional to the level of risk an AI system poses to individuals and society. Since entering into force in August 2024, its provisions have been phased in gradually, allowing organisations time to prepare for compliance.
What changes from August 2026?
As of 2 August 2026, providers of general-purpose AI models, including foundation models powering chatbots, content generation and other AI applications, must comply with a new set of obligations. These include greater transparency regarding model capabilities, technical documentation, and copyright-related information, alongside additional requirements for models that may pose systemic risks. The European Commission, through its AI Office, also gains stronger enforcement powers, including the ability to investigate providers and impose sanctions where necessary.
At the same time, users across Europe will begin noticing one of the Act’s most visible changes: organisations deploying AI must clearly inform people when they are interacting with AI systems or consuming AI-generated content in situations covered by the regulation. Whether through AI assistants, synthetic images, audio or video, transparency is becoming a core principle of responsible AI deployment.
Building trust through transparency
While the new requirements introduce additional compliance responsibilities for organisations, they also provide an opportunity to strengthen confidence in AI technologies.
Transparency is not simply a regulatory obligation; it is a key enabler of trustworthy AI. When users understand when AI is being used, how decisions are supported, and what limitations a system may have, they are better equipped to interact with these technologies confidently and responsibly.
This philosophy closely aligns with the objectives of the AIXPERT project.
Beyond compliance: putting ethics into practice
AIXPERT is developing technologies that make AI systems more transparent, explainable, and human-centric. By combining explainable AI techniques with methods for monitoring trustworthiness and supporting human oversight, the project contributes to an AI ecosystem that reflects both technological excellence and European values. Specifically, designing-in transparency and human-centricity, supported by the project-developed multi-dimensional auditing and evaluation framework, AIXPERT does not only promote compliance with the AI act but fundamentally also alignment with ethical considerations and human values along the whole lifecycle of AI-enabled systems. This includes due care through methods, tools, and best practices from design and implementation, through deployment and operationalisation, all the way to end of life handling for auditable AI outcomes, even when AI systems are taken off from operation.
Looking ahead
The AI Act will continue to roll out over the coming months and years, with further obligations applying to high-risk AI systems and additional governance mechanisms becoming operational. As Europe’s AI regulatory framework matures, collaboration between researchers, industry and policymakers will remain essential to ensuring that innovation develops alongside robust safeguards.
At AIXPERT, we will continue contributing to this vision by advancing research and practical solutions that help make trustworthy AI a reality.