The EU AI Act: From Omnibus to Office – Addressing Risk Across Sectors

On Tuesday, 20 May 2025, SME Connect organized a working meeting titled “The EU AI Act: From Omnibus to Office – Addressing Risk Across Sectors”, hosted by BRANDO BENIFEI MEP, Committee on the Internal Market and Consumer Protection (IMCO) and Committee on Legal Affairs (JURI); Co-Rapporteur on the AI Act (2021–2023), in the European Parliament, Brussels.

The panel featured distinguished experts including KATINKA CLAUSDATTER WORSØE, Digitalisation Attaché, Permanent Representation of Denmark to the EU; BONIFACE DE CHAMPRIS, Senior Policy Manager, CCIA Europe; ANNIKKA LEMMENS, Policy Analyst, OECD; and FRANK MELIS, Founder and CEO of SMRT.bio. The discussion was moderated by JOSIANNE CUTAJAR, Former Member of the European Parliament (2019–2024); Rapporteur for the Opinion on the AI Act (TRAN); Co-Chair of SME Connect’s Working Group on Artificial Intelligence, Data Economy and Cybersecurity; and SME Connect Special Advisor for AI.

JOSIANNE CUTAJAR opened the session by framing the central tension of the AI Act: how to balance innovation with workers’ rights. She posed key questions for the panel, including what qualifies as “high-risk” AI in workplaces, how SMEs can navigate compliance, and how to ensure AI contributes to a fair and safe work environment. Referencing the World Economic Forum and PwC report (November 2024), she highlighted that successful deployment of generative AI depends more on people than on the technology itself – emphasizing the need for training, support, and a cultural shift in organizations. Cutajar stressed that alongside education, regulation must support ethical principles and empower SMEs to innovate and compete. She noted that although the AI Civil Liability Directive has been withdrawn, the European Parliament remains committed to shaping policies on AI in the workplace. Finally, she underlined that both transatlantic dynamics and Europe’s internal political landscape will shape how AI regulation evolves.

BONIFACE DE CHAMPRIS emphasized that artificial intelligence is no longer optional but essential for Europe’s competitiveness. He highlighted that companies using AI effectively report up to a 30% boost in operational efficiency, stressing the need to balance risk discussions with recognition of these benefits. However, he noted a significant AI adoption gap in Europe, with only 13.5% of companies currently using AI – far below the EU’s 75% target. De Champris identified three persistent challenges across company sizes: regulatory complexity, insufficient investment, and data and talent shortages. While he welcomed recent moves by the European Commission to simplify regulation and promote competitiveness, he expressed concern about the low level of ambition and the over-inclusive classification of high-risk systems in the AI Act. He warned that many workplace AI tools – such as those for recruitment, task allocation, and employee evaluation – might fall under strict compliance regimes, depending on future interpretations. Finally, he urged the EU to focus on consistent implementation, adequate enforcement, and transatlantic cooperation, cautioning against protectionism and advocating for a stable, innovation-friendly regulatory environment.

FRANK MELIS introduced his company, SMRT.bio, and its mission helping individuals grow so that regions and entire economies can grow in turn – focusing on psychometric-based profiling, matched with work, learning, and entrepreneurship opportunities. His platform currently supports millions of users, including students and micro-entrepreneurs across Europe, Africa, and soon India, with remarkable engagement and expansion. Melis emphasized the critical role of AI in driving scalable social and economic impact, particularly for underserved communities and micro SMEs, which make up a large share of Europe’s labor force. However, he raised concerns about the complexity and ambiguity of current and upcoming EU AI regulations, especially for small companies like his, asking when and how compliance investments should be made. While acknowledging the value of trust-based European frameworks like the AI Act and GDPR, he warned that they must be balanced with business realities and growth incentives. Ultimately, he urged the EU to better align its regulatory efforts with entrepreneurial needs and platform development, or risk falling behind in global innovation.

ANNIKKA LEMMENS highlighted the dual nature of AI in the labor market – its potential to support workers’ rights, and the significant risks it can pose, particularly through algorithmic management systems. Her OECD research, based on a survey of 6,000 managers, revealed widespread use of such tools, with 78% of EU managers reporting use, but also widespread concern: many managers struggled with accountability and understanding how automated decisions were made. Lemmens stressed that this lack of transparency has real consequences for workers, especially in hiring, training, and contract decisions, and called for clearer company-level procedures and stronger enforcement of existing regulations. She also outlined the positive potential of well-designed AI systems – such as improving work planning, detecting discrimination, and making managerial processes more efficient – if companies engage workers and use explainable tools. Her upcoming OECD study shows 31% of SMEs already use generative AI, mainly for routine tasks, but often without proper training, guidelines, or awareness of legal risks. Lemmens raised specific gender concerns: women are underrepresented in AI-related jobs, more vulnerable to automation due to job type, and less optimistic about AI’s impact. She concluded by calling for gender-sensitive AI policies, inclusive upskilling initiatives, and a focus on coordinated international implementation.

KATINKA CLAUSDATTER WORSØE emphasized a strong national and EU-level focus on simplifying and implementing existing AI rules, rather than creating new ones. She described the current EU regulatory landscape as a “transparent labyrinth” calling for clearer guidance, better explanation of existing tools, and stronger communication with companies – particularly SMEs. Denmark’s upcoming EU presidency will prioritize digital funding and the AI strategy, stressing that underinvestment in digital innovation has contributed to lagging productivity. She noted that enforcing the AI Act, Product Liability Directive (PLD), and Platform Work Directive is more urgent than introducing additional laws, and warned against adding unnecessary regulation that creates more red tape for businesses. Worsøe highlighted the importance of standard-setting – especially for high-risk AI – and urged industry stakeholders to participate actively, as these standards are being written now. She also stressed the need for collaborative governance, particularly between the AI Office, AI Board, member states, and other actors, to provide timely and useful guidance. Concluding, she called for an evidence-based approach to any future legal changes: enforce what’s already been adopted before expanding the regulatory framework.

BRANDO BENIFEI MEP emphasized that the EU AI Act aims to increase, not limit, the uptake of AI by fostering trust and responsibility in development. He underscored the need to protect SMEs from bearing undue regulatory burdens while holding large AI developers accountable. Acknowledging early implementation challenges, he warned of emerging oligopolistic dynamics and stressed the need for fair competition and support for European AI startups. Benifei called for complementary measures beyond regulation, including skills training, lifelong learning, and better digital infrastructure. He urged further development of AI and GenAI factories, deeper capital markets, and removing structural investment barriers for EU startups. While expressing regret over the political barriers to a dedicated AI-in-the-workplace directive, he affirmed continued efforts to strengthen the AI liability regime and avoid legal fragmentation across member states. He closed with optimism about finding pragmatic compromises in the legislative process and reaffirmed support for a balanced, innovation-friendly AI framework in Europe.