Impact of AI on Romanian Businesses (survey dataset description)

Impact of AI on Romanian Businesses (survey dataset description) BACKGROUND AND RATIONALE Rapid advances in artificial intelligence (AI) have prompted growing interest in how these technologies affect businesses worldwide. However, evidence from recent statistics indicates that AI adoption in Romania remains nascent. In 2023, only about 1.5% of Romanian enterprises (with ≥10 employees) had implemented AI solutions – the lowest share in the EU, compared to roughly 8% EU-wide (Eurostat, 2024)[1][2]. This gap underscores the importance of understanding local drivers and barriers to AI uptake. Romanian policymakers have also emphasized the need for strategic action: a National AI Strategy was launched to maximize the benefits of AI integration and align with European trends (Autoritatea pentru Digitalizarea României & MCID, 2024). Within the private sector, there is a mix of enthusiasm and caution. For instance, a recent entrepreneurship survey found 71% of Romanian firms had increased their use of AI in the past year (EY Romania, 2025)[3], yet 37% reported challenges in training employees on AI tools[4]. In this context, the present survey was conceived to systematically assess the impact of AI on businesses in Romania, filling an empirical gap by capturing business stakeholders’ experiences, readiness, and concerns regarding AI. The survey addresses a wide array of themes – from workforce implications and cybersecurity challenges to changes in financial services, ethical considerations, and government policy related to AI – reflecting the multifaceted nature of AI’s business impact[5]. The primary rationale for the project was to gather robust data that can inform both academic analysis and practical strategies for AI adoption in the Romanian business environment. SURVEY STRUCTURE AND INSTRUMENT The “Impactul Inteligenței Artificiale asupra Afacerilor din România” survey is a structured questionnaire developed in Romanian language using the LimeSurvey online platform. It was designed as a self-administered web survey targeting individuals knowledgeable about their organisation’s use of AI (e.g. executives, managers, IT specialists, entrepreneurs). The instrument comprises 20 questions organised into nine thematic sections, ensuring coverage of both contextual profile information and specific AI-related topics. Below is an overview of the survey structure: · Section I – Organisational Profile: 3 questions on company size, industry sector, and the respondent’s role in the company. · Section II – AI Familiarity & Adoption Motives: 2 questions assessing the respondent’s familiarity with AI applications in business and the primary motives for adopting AI solutions. · Section III – AI Adoption Barriers: 1 multi-select question identifying the main perceived obstacles to implementing AI in the organisation. · Section IV – AI and Employment: 2 questions on how AI is expected to affect jobs (e.g. job creation vs. replacement) and what measures the company is taking (or would take) to train or prepare employees for AI integration. · Section V – AI in Business Future: 2 questions examining anticipated high-impact AI application areas for business value and opinions on AI’s potential to drive economic growth. · Section VI – Regulation & Ethics: 2 questions regarding the need for AI-specific regulation to mitigate risks and the primary ethical concerns related to AI use in business. · Section VII – Global Competitiveness: 2 questions gauging perceptions of Romania’s competitiveness in AI adoption relative to other European countries and how Romania should position itself in the global AI market. · Section VIII – Emerging Topics & Strategy: 5 questions covering strategic outlooks, including AI’s environmental impact, the role of government incentives in AI adoption, expected changes to the company’s business model due to AI, anticipated new AI innovations in the industry (open-ended), and approaches to measuring return on investment (ROI) for AI projects. · Section IX – Final Considerations: 1 open-ended question inviting any final comments or suggestions from respondents regarding AI’s impact on their business or the Romanian market. This survey structure was purposefully designed to capture both quantitative inputs (through closed-ended questions in single-choice or multiple-choice format) and qualitative insights (through two open-ended questions in Sections VIII and IX). The closed questions use categorical response scales tailored to each topic (including ordinal scales for level-of-familiarity and agreement where appropriate), while multiple-selection questions allow respondents to choose all applicable options (e.g. selecting all relevant barriers or training measures). The instrument underwent content validation to ensure clarity in Romanian business context, and it was estimated to require ~10 minutes to complete, balancing depth with respondent effort[6]. The overall purpose of the survey instrument is to obtain a comprehensive yet focused snapshot of how AI is perceived, adopted, and managed in Romanian businesses, providing a basis for descriptive analysis without delving into individual companies’ proprietary data. DATASET DESCRIPTION The responses collected from this survey have been compiled into an anonymised dataset (in CSV format) for research and archival use. The dataset captures each respondent’s answers across all questions, with one row per completed response. Table 1 summarises the key characteristics of the dataset and survey content: DATASET ATTRIBUTE DESCRIPTION Total responses 212 completed responses (unique organisations/respondents) Survey period 2024–2025 (data collected in this interval) Total questions 20 questions, structured into 9 thematic sections (as detailed above) Question formats 13 single-choice categorical questions; 6 multiple-selection questions; 1 open-ended question Total variables (columns) 139 columns in the CSV dataset, including all survey question fields and sub-items for multi-select options, plus technical metadata Thematic coverage Organisational demographics; AI familiarity and usage; drivers and barriers to AI adoption; workforce impact; strategic outlook on AI (business value, competitiveness, innovation); regulatory and ethical perspectives; and open-ended feedback Each row in the dataset corresponds to an individual survey submission, identified by a unique response ID. The 139 columns encompass: system fields (e.g. response ID and timestamps); one column for each single-choice question; multiple binary or categorical columns for each multi-select question (indicating which options were chosen, with accompanying free-text columns for any specified “Other” inputs); and text fields for the open-ended questions. All variable names are coded for brevity in English or alphanumeric identifiers, while the survey questions were administered in Romanian. A codebook is provided along with the data, mapping each column to its survey question and response options. In terms of data types: most variables are nominal or ordinal categorical data (recorded as numeric codes or text labels for selected categories), and a smaller number are free text strings (for the open-ended responses). There are no continuous numerical measurements in this survey; rather, the emphasis is on categorical perception and decision data. The structure enables researchers to perform frequency analysis, cross-tabulations (e.g. comparing responses by company size or industry), and qualitative content analysis on open-ended answers, without needing any personal identifiers. Notably, the dataset reflects a diverse range of company profiles and AI engagement levels due to the broad outreach of the survey. While this documentation does not interpret the results, basic counts indicate that the 212 respondents cover various industries and organizational sizes, making the dataset a valuable cross-section of Romanian business perspectives on AI. The multi-topic coverage, from cybersecurity concerns to ROI evaluation, allows for multifaceted descriptive analysis. For instance, researchers can identify how many companies cite high implementation cost as a barrier versus those that cite lack of expertise (from the multiple-choice barriers question), or enumerate the kinds of AI applications that respondents consider most valuable to their business (from the multi-select question on high-potential AI applications). DATA ANONYMISATION AND ACCESS This dataset has been thoroughly anonymised to protect respondent privacy and confidentiality. No personally identifiable information (PII) was collected in the survey instrument – questions focused on organisational characteristics and attitudes rather than individual identities, and the survey did not ask for names, contact details, or company names. Furthermore, any indirect identifiers (such as IP addresses or precise timestamps) have been removed or generalized in the public dataset. All responses are stored under randomised ID codes. As a result, individual respondents or organisations cannot be identified from the data, in compliance with research ethics and data protection standards. The full survey questionnaire remains available online for consultation at the original LimeSurvey portal (in Romanian) at https://www.info-world.ro/so/ index.php/444639?lang=ro . This ensures transparency of the instrument and allows interested parties to view the exact phrasing of questions and options as presented to participants. The online availability of the survey also enables ongoing data collection; however, the dataset described here corresponds to a fixed export of completed responses as of the survey’s initial analysis period. Researchers and practitioners are welcome to use this dataset under the terms of its open data license, and to cite this documentation report in any derived publications. By providing a detailed description of the survey instrument and dataset, we aim to facilitate the reusability of the data and encourage further scholarly examination of AI’s business impact in emerging markets such as Romania. REFERENCES Autoritatea pentru Digitalizarea României (ADR) & Ministerul Cercetării, Inovării și Digitalizării (MCID). (2024). Strategia națională în domeniul inteligenței artificiale 2024–2027 (National AI Strategy 2024–2027). Government of Romania. Eurostat. (2024, May 29). 8% of EU enterprises used AI technologies in 2023 [News release]. Eurostat Statistics Explained. [1][2] EY Romania. (2025, May 29). EY Entrepreneurship Barometer Study: 71% of companies adopt Artificial Intelligence, while 75% fear economic risks in Romania [Press release]. EY.com. [3][4] [1] [2] 8% of EU enterprises used AI technologies in 2023 - News articles - Eurostat https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20240529-2 [3] [4] EY Entrepreneurship Barometer Study: 71% of companies adopt Artificial Intelligence, while 75% fear economic risks in Romania | EY - Global https://www.ey.com/en_ro/newsroom/2025/050/ey-entrepreneurship-barometer-study- [5] [6] INFLUENȚA INTELIGENȚEI ARTIFICIALE ÎN MEDIUL DE AFACERI DIN ROMÂNIA https://surveys.tmedia.ro/index.php/854239?lang=en