CEE’s Pro-Growth AI Agenda

Central and Eastern Europe should approach AI policymaking from a realistic starting point: Europe has zero near-term ability to match the United States or China in the most capital-intensive parts of frontier AI development. The latest Stanford AI Index Report shows that U.S. private AI investment reached $285.9 billion in 2025, more than 23 times China’s $12.4 billion. U.S. entrepreneurial activity also remained far ahead, with 1,953 newly funded AI companies in 2025, more than ten times the next closest country. Stanford also reports that the U.S. and China have traded places at the top of frontier model performance rankings since early 2025, with the performance gap effectively closing, while industry produced over 90% of notable AI models in 2025.

The AI investment gap is part of a deeper competitiveness crisis. EU labour productivity has fallen below 80% of U.S. levels, and the critical difference is the technology sector. Europe's share of global AI patents has plummeted from around 10% in 2010 to less than 3% today. Research published by the National Bureau of Economic Research found that U.S. venture capital investments in Europe dropped more than 20% following the implementation of the GDPR. The European Commission's own 2025 Digital Decade Report acknowledged that "several key domains, particularly foundational digital technologies such as AI, cloud services, data analytics, are showing deeply unsatisfactory progress." Meanwhile, 61% of Single Market barriers identified in 2002 still existed in 2020.

This does not mean Europe, or CEE, should lower its AI ambition. It means the region must compete differently. If Europe cannot match the U.S. and China on venture capital, hyperscale infrastructure, public investment, large domestic markets or frontier-model ecosystems, it must use the policy levers it can control. The most important of these is the regulatory and legislative environment. Europe should become the easiest major democratic market in which to research, test, deploy, commercialise and scale responsible AI.

Yet Europe is moving in the opposite direction. The AI Act now sits alongside the GDPR, ePrivacy rules, cybersecurity obligations, product-safety legislation, data rules, platform regulation, consumer law and sector-specific regimes. For CEE startups, SMEs and researchers, this cumulative rulebook can determine whether AI systems are developed and scaled in Europe at all. 

The message from industry is unequivocal. In a joint “Europe’s Tech Creators’ Wake-Up Call” the CEOs of Airbus, ASML, Ericsson, Mistral AI, Nokia, SAP and Siemens warned that Europe faces “stifling, unnecessarily complex and often overlapping rules” and called for digital rules to be reduced and simplified so they serve as “agile guardrails” rather than rigid requirements. The European Tech Alliance (EUTA) has called on the EU to “scrap obsolete rules, stop layering new ones, and deliver smarter regulation that unlocks growth.” Twenty Member States have written to the European Council President demanding a systematic review of all EU regulations to identify rules that are “superfluous, excessive, or unbalanced.” 

CEE should therefore advocate a pro-growth and deregulatory AI agenda. This does not mean deregulation for its own sake. It means removing, narrowing, consolidating or repealing rules that are duplicative, disproportionate, unclear, outdated or hostile to innovation. Europe should maintain safeguards where they address real risks, but stop layering obligations that create uncertainty without delivering meaningful protection.

A pro-growth AI agenda for CEE should rest on three priorities: simplifying and deregulating the AI rulebook; completing the Single Market for AI; and boosting practical CEE AI cooperation.

1. Simplify and deregulate the AI rulebook

Regulatory simplification must become the centrepiece of Europe’s AI competitiveness agenda. Europe cannot outspend the U.S. or China in the short term, but it can decide whether its rules help or hinder responsible AI development.

The AI and Digital Omnibus should be treated as a turning point. Recent trilogue progress shows that EU institutions recognise the need for clearer timelines, lower administrative burdens and more practical rules for companies embedding AI into products and services. This direction must now be strengthened, not diluted.

Europe should preserve pro-innovation clarifications on personal data, pseudonymisation, scientific research, legitimate interest for AI training and incidental sensitive data. Without clear rules in these areas, companies will face higher compliance costs, slower time-to-market and weaker incentives to invest in Europe.

Germany has taken a constructive lead by supporting a simpler, more innovation-friendly and business-oriented EU digital framework. Central and Eastern Europe should follow this lead. CEE governments should actively push for a smarter AI rulebook that protects citizens while enabling startups, SMEs and industrial companies to build, test and scale AI in Europe.

The message should be clear: simplify the rulebook, reduce legal uncertainty and make Europe a place where responsible AI can grow.

CEE governments should act on four commitments.

First, recognise and operationalise the right to innovate as a core European principle. The EU should treat the ability to research, develop, test and commercialise new technologies as a core condition of a competitive and open society. Innovation enables better healthcare, stronger education, more effective public services, productivity growth and new jobs. Regulators should therefore not only assess risks, but also consider whether their decisions support research, investment, productivity and the freedom to conduct business.

Second, require serious impact assessments before new rules are adopted. The OECD, in its Better Regulation Practices across the European Union 2025, states that rigorous impact assessments improve economic welfare. Yet the same report also finds that the EU’s current approach to impact assessments is itself a driver of “persistent” regulatory burden. No new AI or digital rule should proceed without evidence of a genuine problem, proportionality and implementation feasibility. Impact assessments should measure effects on innovation, research, AI training, startups, SMEs, public administrations, investment, time-to-market and cross-border scaling. They should also assess cumulative burden across the full digital rulebook, not only the marginal cost of each individual file. If the Council or Parliament introduces major amendments, the assessment should be updated before adoption.

Impact assessment should not stop once legislation is passed. Major AI and digital rules should include ex-post evaluation, sunset review or mandatory revision clauses. If a rule fails to achieve its objective, creates disproportionate costs or harms innovation without clear benefits, it should be simplified, narrowed or repealed. For regulatory non-compliance decisions impacting the single market, regulators should also conduct impact assessments that consider effects on consumers and innovation.

Third, make innovation enforceable and implementation simple. Europe has strong mechanisms to enforce compliance and restrictions, but too few mechanisms to protect innovation, research and competitiveness in practice. Regulators should be required to show how their guidance and enforcement support innovation as well as protection. The Commission should maintain a live overview of cumulative regulatory burdens across the digital rulebook and identify overlaps, conflicts and disproportionate costs.

The same discipline should apply nationally. CEE governments should avoid gold-plating EU AI and digital rules through extra registers, authorisations, reporting duties, fees or unclear supervisory layers. Each country should publish plain-language guidance, standard templates, one-stop contact points, clear regulator responsibilities and examples of compliant low-risk AI use cases. CEE governments should treat startups and scale-ups as a distinct implementation category, with guidance and support adapted to their limited compliance capacity and need to scale quickly. Where AI rules overlap with data protection, cybersecurity, product safety or consumer law, national regulators should issue one clear interpretation rather than leaving companies to reconcile conflicting requirements.

Fourth, pause new broad digital regulation and focus on simplification. There is no logic in adding new horizontal rules while simultaneously trying to simplify existing ones. Europe should halt new broad digital regulation until the current rulebook has been reviewed and should prioritise repeal, consolidation, clearer legal bases for AI training and research, and removal of duplicative documentation before adding new horizontal digital rules. The Digital Omnibus should be treated as the beginning of a wider regulatory reset, not as a one-off technical correction. CEE governments should use the Digital Fitness Check to identify rules that are obsolete, overlapping, disproportionate or harmful to innovation.

To make this agenda effective, CEE governments should coordinate common Brussels positions early. Digital ministries, permanent representations and national regulators should align on shared priorities: no gold-plating, simple implementation, clearer rules for AI training and research, stronger impact assessments and genuine simplification.

2. Complete the Single Market for AI

Europe has the market size to compete in AI, but companies do not experience it as one market. They face different regulators, diverging interpretations of EU rules, fragmented procurement systems, inconsistent data-access practices and different administrative expectations. For CEE, this fragmentation is especially damaging. Startups, scale-ups and research spinouts from smaller markets must scale across borders early, but too often they are forced to treat each Member State as a separate regulatory and commercial environment.

CEE governments should push for an immediate Single Market agenda for AI focused on practical market access.

First, AI companies should face one clear regulatory pathway across Europe. Where the AI Act overlaps with data protection, cybersecurity, consumer protection, product safety or sectoral rules, regulators should coordinate and issue joint guidance. A company should not receive 27 different interpretations of the same EU obligation.

Second, Europe should move toward mutual recognition of regulatory guidance, sandbox outcomes and conformity support. If a startup tests an AI product in one Member State or receives structured guidance from a competent authority, that outcome should carry weight across the Single Market unless the product or use case materially changes. This would reduce duplication and make cross-border scaling realistic for CEE firms.

Third, the EU should deliver a startup-friendly legal environment. CEE governments should support the 28th regime / EU Inc. agenda: an optional, simple EU-wide company framework for innovative firms operating across borders. This should make incorporation, employee stock options, insolvency, fundraising and cross-border operations easier for startups and scale-ups.

Fourth, public procurement should become a Single Market tool. Public buyers should avoid national preferences, excessive turnover requirements and tender conditions designed only for large incumbents. Procurement should open startup access through smaller lots, proportionate requirements and scalable pilot contracts. CEE governments should promote modular tenders, challenge-based procurement and common AI procurement clauses.

Europe should also apply a smart “buy European” principle where credible European alternatives exist, especially in sensitive areas such as public administration, cybersecurity, health, defence-adjacent technologies and critical infrastructure. This should not become protectionism. In many parts of the AI stack, including cloud, compute, foundation models and developer tools, cooperation with U.S. technology vendors will remain necessary because Europe does not yet have equivalent alternatives at scale. Premature mandates in areas where no European alternative exists would effectively mean "buy nothing" or "build worse", harming competitiveness rather than helping it. CEE governments especially depend on access to global AI tools to accelerate digital transformation given smaller national budgets. The right approach is “European where possible, best available where necessary, and open to trusted partners where dependencies cannot yet be avoided.”

Finally, data access should support cross-border AI development. CEE governments should make public-sector datasets easier to access, standardise formats and licensing, and support secure cross-border data-sharing projects in areas such as health, transport, energy, language technologies, public administration and fraud detection.

3. Boost practical CEE AI cooperation

CEE countries should cooperate on AI in a realistic and practical way. The region does not need heavy new institutions, unrealistic joint funds or another broad strategy with dozens of priorities. It needs disciplined coordination, faster exchange of know-how, stronger representation in Brussels and a small number of practical cross-border projects where cooperation clearly adds value.

First, CEE governments should coordinate common AI positions in Brussels. Digital ministries, permanent representations and digital attachés should meet ahead of major Council negotiations on AI, data, cybersecurity, digital regulation and the Single Market. The goal is not a formal bloc, but earlier alignment on shared priorities: simplification, no gold-plating, legal certainty for AI training and research, startup-friendly rules, open procurement and stronger Single Market enforcement. Beyond individual files, CEE should jointly advocate for structural reforms at EU level, including mandatory international.

competitiveness benchmarks in all impact assessments, sunset clauses in major digital legislation, and institutional accountability for the simplification agenda.

Second, CEE countries should share implementation know-how. Each government will need to implement the AI Act, prepare guidance, train regulators, support sandboxes, advise companies and modernise procurement. Instead of reinventing templates, CEE administrations should exchange model guidance, risk-assessment forms, procurement clauses, sandbox procedures, public-sector AI playbooks and examples of compliant low-risk use cases.

Third, CEE should build public-sector AI capacity together. The region faces similar shortages of AI, IT, procurement and digital-policy expertise in public administration. Governments should organise practical training for civil servants, regulators, procurement officials, diplomats, municipal leaders and digital attachés: how to procure AI, assess risk, avoid overimplementation and represent national AI interests in Brussels.

Fourth, CEE governments should launch a small number of cross-border AI public-service pilots. Cooperation should focus on use cases where scale and shared data create clear benefits, e.g. AI translation for smaller European languages, customs and fraud detection, cybersecurity threat sharing, transport data, EU funds control, disaster response, or interoperable digital identity support.

Finally, CEE should improve regional visibility for AI startups, research teams and public-sector innovation. Governments and innovation agencies should coordinate representation at major European and global AI events, investor missions and policy forums. CEE should be more visible as a source of AI talent, applied research, industrial AI solutions and trustworthy public-sector innovation.

CEE countries cannot afford to wait for Europe’s AI position to improve by itself. The call to action is clear: coordinate early in Brussels, push for real deregulation and simplification, stop gold-plating at home, make the Single Market work for AI companies, and cooperate practically where scale matters. Europe must reconnect with its roots as an economic and industrial powerhouse.

Europe will not become an AI leader by adding complexity. CEE should lead the case for a different path: clear rules, open markets, faster deployment, stronger research commercialisation and practical cooperation. The region has the talent, the reform mindset and the political will to lead Europe's regulatory reset. This is how the region can move from AI potential to AI impact.

Respectfully,

🔹Michal Kardoš, Executive Director, SAPIE - Slovak Alliance for Innovation Economy

🔹Jaromír Hanzal, Director, AAVIT - Association for Applied Research in IT 

🔹Tomasz Snażyk, CEO, AI Chamber CEE

🔹Corina Vasile, Executive Director, ANIS - Employers’ Association of the Software Services Industry

🔹Róbert Gašparovič, Chairman of the Board, ASAI - AI Association

🔹Andrei Kelemen, CEO, Cluj IT Cluster

🔹Egle Markeviciute, EU Affairs Manager, Consumer Choice Center Europe

🔹Marijana Šarolić Robić, Vice President, CRO STARTUP

🔹Tomáš Pospíšil, Chief Analyst, Czech Startup Association

🔹Michal Kanownik, President, Digital Poland Association

🔹Nenad Sutanovac, Director, ICT Association of Slovenia

🔹Simonas Černiauskas, CEO, Infobalt

🔹Doris Põld, Executive Director, ITL - Estonian Association of Information Technology and Telecommunications

🔹Mária Bieliková, Director General, KInIT - Kempelen Institute of Intelligent Technologies

🔹Lenka Kučerová, Director, PragueAI

🔹Matej Rus, Leader, Startup Slovenia