Artificial intelligence (AI) has moved from a technological opportunity to a strategic imperative for Hungarian and European firms, according to a panel on AI at the 64th Közgazdász‑vándorgyűlés. Panelists agreed that AI can materially boost productivity, but warned that technology by itself does not ensure success: organisational change, deep domain expertise, software and viable business models are also required.
Differences between SMEs and larger companies
Vilmos Levente Kovács, CEO of Simplexion Informatikai Kft., described how his company has been developing AI‑assisted solutions for about one and a half years. He said developers increasingly act as system organisers rather than writing code directly, and his firm currently works on three products built this way. Kovács argued that competitive advantage today comes from specialised domain knowledge, so for small firms investment in knowledge is often more important than pure capital expenditure.
By contrast, Balázs Veszprémi, CEO of ARTEMIS Technologies Zrt., warned that industrial‑scale systems require significant investment and scaling capacity. His roughly twenty‑person development team builds IT systems that, among other things, manage and optimise power‑plant operations second‑by‑second for major energy players. Here too domain expertise is increasingly valuable, while it becomes harder to find professionals who combine technological and industry competencies.
Size, adoption and labour market effects
Benedek Lits, a researcher at Budapesti Corvinus Egyetem and the lead on the forthcoming Eurochambres Working Paper, noted that AI adoption correlates with firm size but a country’s lagging position is not simply explained by a high SME share. Their European data set found no substantive relationship between employment share of medium‑sized firms and AI use. Their regression analysis indicates that certain AI uses — for example process optimisation and some image and data processing applications — are positively associated with firm performance.
Lits also warned that generative AI is already producing detectable labour‑market effects: European employment data show that after ChatGPT appeared in 2022, sectors more exposed to generative AI experienced a significant decline in employment of workers under 25.
Training and adaptation examples
Lits highlighted European programmes preparing young people for AI: for instance, Germany offers an accredited AI qualification tied to 1,250 hours of practical experience, and other initiatives train youths to deploy AI solutions in SMEs.
Europe’s strategic dependence and sovereignty
Antal Kuthy, co‑founder and CEO of E‑Group ICT Software Zrt., framed Europe’s situation as a strategic problem: he argued the continent lost the first phase of the digital race and became dependent across several key technology layers. He pointed out that approximately 85 percent of European data‑center capacity is in American hands, and significant American ownership is present in much of the remaining capacity. Kuthy said technological sovereignty is no longer an industrial policy buzzword but an issue of economic independence.
Kuthy and Pongrácz Ferenc, the company’s CFO, demonstrated a sovereign AI system running in Hungary. The solution runs in a local data centre and does not send corporate or state data to foreign clouds. In the demonstration the system first flagged an energy‑market risk following a Brent oil price increase and recalculated budgetary effects: an initial 2.9 percent deficit path would rise to 3.3 percent without intervention. The AI modelled changes in revenue and expenditure channels, projected social impacts, prepared alternative policy packages and produced a draft government submission, implementation tasks and a communications plan.
Kuthy argued the main value of such systems is keeping strategic data and decision processes out of external providers’ hands.
Return on investment and the need for measured decisions
Magdolna Csath, chair of the Innovation Section of the Magyar Közgazdasági Társaság (MKT), emphasised that companies must evaluate returns: they need to define the problem they will solve with AI, the value it can create, the required investment and the expected payback. Citing a corporate survey, she noted that about 70 percent of companies have reached a point where employees use AI individually, while only 27 percent have done so at an organisational level.
Csath used international indicators to illustrate Hungary’s position: the 2026 IMD competitiveness survey places Hungary 57th out of 70 countries for skills required for AI. One dataset measuring corporate digitalisation found only 6.2 percent of Hungarian firms highly digitalised, while 39 percent were very low in digitalisation. She also highlighted that AI use is particularly low among companies with 10–49 employees.
Csath argued AI should be used to create new products, services and markets and workers should be retrained; otherwise expensive technologies risk leading to excessive investment, higher indebtedness and larger regional disparities.
Debate over dependence on external providers and the EU’s role
A sharp debate focused on the risks of reliance on external AI providers. Kovács said many SMEs can use AI with no or low upfront investment, citing a camera‑based person‑recognition solution that runs on a roughly HUF 100,000 local device with an open model, keeping personal data off the cloud. Veszprémi argued that much corporate data is not so sensitive as to require local handling, so risk assessment must be use‑case specific.
Kuthy warned the issue is broader than data protection: firms can create strategic dependency if they build critical processes on an external platform that might later raise prices or offer the same capability as its own product.
Panelists also disagreed on the role the state and the European Union should play in financing technological catch‑up. Kuthy noted EU initiatives recognise market failures in certain areas and that private capital alone may not close the gap. Csath cautioned that despite substantial prior EU funding, results are often lacking: Europe still trails China and the United States in patents for new technologies, and many promising companies grow large in the United States.
Conclusion: specialise and reduce strategic dependency
The discussion concluded that neither Hungary nor Europe should aim to replicate the largest American or Chinese models. A more realistic path to competitiveness is building specialised, domain‑knowledge applications, accelerating corporate adaptation and developing the technology layers that reduce strategic dependence.
