By Alycia Colijn
In September and October of 2025, the European Commission called on European citizens and organizations for evidence on their Digital Omnibus initiative, an attempt to simplify rules around digital innovation in order to foster innovation. Below, you can find our response to the call. On November 19, the Commission subsequently published their plans for the Digital Omnibus.
Simplification should however not lead to weakening of regulatory safeguards that are in place. The AI Act is complex structurally and linguistically, thus an attempt at simplifying for efficiency and effectiveness is welcome but needs careful implementation strategies so as not to lose precision and specificity and cause ambiguity, or less stringent protection of human rights. We suggest careful legal drafting and strong inter-institutional cooperation to avoid losing the regulatory basis of such Acts due to simplification.
In striving for simplification, Responsible AI should not be governed by a culture of tick-the-box ethics. That brings us to a few remarks.
Firstly, most entrepreneurs struggle with the complexity of rules, as well as the question of which rules apply. Therefore, we believe a solid communication strategy (possibly in collaboration with national compliance authorities) is a vital part of simplification. These could include visual product-launch journeys that include details on both the AI Act and the subsequent guidelines on prohibited and regulated systems that have been introduced.
Secondly, we encourage a sector-specific digital omnibus and suggest that the recently announced AI Act Advisory Forum sets up a working group around simplification, including representatives of advocates for entrepreneurial affairs (e.g. Chambers of Commerce and advocacy organizations).
Thirdly, we recommend specific guidelines for software developers and a question list that entrepreneurs can use when working with software partners. Most SMEs will not develop advanced AI models but rely on software embedding AI or build on foundational models. While the General Purpose AI Code of Practice guides the foundational models, tailored guidance for developers would facilitate AI Act adoption.
Fourthly, we believe the striving for simplification should not just be an ambition of administrative simplification, but an attempt to help small-cap entrepreneurs with the ethical implementation of AI in their business models, as most would not have the resources to conduct holistic risk analyses. Therefore, we argue for standardized, easy-to-access-and-use ethical tools, such as: open-access AI-ethics impact templates, shared databases with best practices, and common audit frameworks within sectors. These resources could be shared via the AI on Demand resource database initiative; however, AIoD would need rigorous simplification to make it user-friendly as well.
Fifthly, for the upcoming Digital Fitness Check of Consumer Laws, we would welcome recommendations in the field of contextual advertising, which reduces pressure on data centers and computing power, improves privacy, and — research shows — leads to higher conversion for commercial parties. It also mitigates monopolization risks in the field of online tracking and consumer data. For the Digital Fitness Check that will analyze the cumulative effects of the simplification, we encourage the Commission to not only analyse the reduced administrative burden, but also include ethical performance indicators, e.g. transparency of AI-related decision-making and notifications of bias and harm caused by AI systems.
Lastly, on data legislation, we encourage the Commission to prioritize, or where possible require, federated learning and on-device processing. These protect privacy, benefit entrepreneurs, and support sustainability by reducing computing power demand, aligning naturally with GDPR objectives. As the AI Act requires non-discrimination, accuracy, and traceability in datasets, we advocate for more explicit provisions ensuring datasets meet these standards. This would accelerate ongoing standardization across sectors and member states.
We deeply appreciate the Commission’s ongoing work to ensure responsible AI development and stands ready to support the implementation.
