ReSTART Knowledge Observatory — Poland

Compiled by WSEI (Uniwersytet WSEI w Lublinie) · Polish partner · October 2026

Evidence base on AI, the labour market and skills in Poland

This report brings together the newest official statistics, research reports and policy documents on artificial intelligence, the labour market and adult skills in Poland, with a focus on the four ReSTART job profiles (Virtual Assistant, Customer Service Agent, Administrative Support Specialist, Social Media Assistant) and on unemployed adults with low to medium digital skills.

Poland at a glance — key figures for ReSTART

AI in enterprises and the workplace

1. Społeczeństwo informacyjne w Polsce w 2025 r. (Information society in Poland in 2025)

Statistics Poland (GUS), Statistical Office in Szczecin · 16 December 2025 (first release 21 October 2025; tables 8 December 2025) · official statistics, 2025 survey

The annual GUS publication reports the results of the EU-harmonised surveys on ICT use in enterprises (10 or more persons employed) and in households (people aged 16–74). It is the official national source on AI adoption and the newest edition available on 1 October 2026.

In 2025, 8.7% of Polish enterprises used at least one AI technology (5.9% in 2024). Use was highest in large enterprises (42.0%) and in information and communication (37.8%), and lowest in small enterprises (6.1%). Generative technologies (text or speech generation) were the most common type, used by 5.3% of enterprises. AI was used most often for marketing and sales (5.0% of all enterprises) and for office work related to HR and business management (3.7%). 91.3% of enterprises did not use AI; the most frequent reason given was a lack of human resources and knowledge about using AI.

Among people aged 16–74, 22.7% used AI tools (such as ChatGPT, Copilot or Gemini) in the three months before the survey, but only 8.4% used them for work. Use was 12.9% among unemployed people, 12.6% among people with secondary education (2.3% for work) and 7.5% among people aged 55–64. Most non-users said they did not need such tools (54.5%); 4.2% pointed to a lack of skills.

For ReSTART: Official data show that AI is still rare in small Polish firms and that unemployed and secondary-educated adults use AI tools far less than others. ReSTART’s target group is on the wrong side of this gap, and many do not yet see why AI matters for them — so courses should start from concrete, job-related benefits.

2. 20% of EU enterprises use AI technologies — and The use of artificial intelligence (AI) technologies in the European Union: Key results, 2026 edition

Eurostat · 11 December 2025 (news) and spring 2026 (statistical report) · official EU statistics, 2025 data

Eurostat’s harmonised enterprise survey allows a direct comparison of Poland with the EU. It covers enterprises with 10 or more employees in the business economy (excluding the financial sector).

In 2025, 20.0% of EU enterprises used AI (13.5% in 2024). Poland, at 8.4%, was among the three lowest countries together with Romania (5.2%) and Bulgaria (8.5%). The gap is widest among small and medium-sized firms: 5.79% of small and 14.77% of medium-sized Polish firms used AI, compared with 17.0% and 30.36% in the EU, while large firms (45.81%) are closer to the EU level (55.03%). Eurostat notes that Poland has one of the largest differences between medium-sized and large firms (31.0 percentage points).

For ReSTART: Most ReSTART graduates will look for work in small and medium-sized firms, where AI adoption in Poland is lowest. A worker who can introduce simple, safe AI routines can bring real value to such employers.

3. Szybki Monitoring NBP. Analiza sytuacji sektora przedsiębiorstw, nr 03/26 (NBP Quick Monitoring Survey, July 2026)

Narodowy Bank Polski (National Bank of Poland) · 20 July 2026 · enterprise survey (Annual Survey: 1,913 firms, April–May 2026)

The central bank’s regular survey of non-financial enterprises included questions on AI in its 2026 annual module (1,913 firms from all sectors except agriculture). NBP notes that large firms are over-represented and micro firms under-represented, so results should be read with caution.

The share of surveyed firms using AI tools rose from 14% in Q4 2024 to 42% in Q2 2026, and 29% increased their use of AI during 2025. Use is higher in large (59%) and medium-sized (43%) firms than in small and micro firms (28%). So far, the effect on jobs has been minimal: fewer than 1% of firms say that wider use of AI allowed them to reduce staff. 35% of firms have not yet analysed how they could use AI.

For ReSTART: Polish evidence from 2026 supports the ReSTART message that AI is changing tasks rather than eliminating jobs: firms are adopting AI quickly, but very few are cutting staff because of it.

4. Sztuczna inteligencja w firmach: gotowość do adopcji, kompetencje i potrzeby. Raport z badań BKL (Artificial intelligence in companies: readiness, competences and needs)

Polish Agency for Enterprise Development (PARP) and Jagiellonian University, Human Capital Study (BKL) · 2026 (survey June–September 2025) · mixed-methods study; survey of 1,822 employers in 12 industries with the highest AI potential

This BKL study combines interviews, expert workshops and a weighted random-sample survey of employers in the 12 industries with the greatest potential for AI. The results are representative of those industries, not of the whole economy.

23% of firms use AI in their business processes and another 13% see potential or plan to adopt it; over 80% are at an early stage of readiness. Customer service is the second most common process where firms use or plan to use AI (34%), mostly for handling customer tickets, complaints, queries and requests (71%). Nearly a third use or plan AI in data and document management (30%), customer service and communication (29%) and marketing (28%). Administrative positions are the most frequently named area of AI use (26%). 64% of AI-using firms rely on free, general-purpose tools, and only 28% have access to qualified IT staff. When skills are missing, firms most often train their current employees (23%).

For ReSTART: The tasks where Polish firms apply AI — customer enquiries and complaints, documents, marketing, administration — are exactly the tasks of the four ReSTART profiles. Because most firms use free tools without IT support, employees need to know how to use them safely and check their results.

5. Badanie Pracuj.pl: 70% pracowników nie korzysta z AI w pracy (Pracuj.pl survey: 70% of workers do not use AI at work) — “To tylko AI” study

Pracuj.pl (Grupa Pracuj), fieldwork by ARC Rynek i Opinia · 16 September 2026 · online survey (CAWI) of 1,506 working Poles, August 2026

The survey by Poland’s largest job portal looks at how working Poles actually use AI. 70% of respondents do not use any AI tools or functions at work. Among those who do, 69% stay with simple commands and prompts. Workers estimate that AI could take over, on average, 21% of their working time, and 23% say that AI has given them more tasks.

For ReSTART: In autumn 2026 most Polish workers still do not use AI at work, and users stay at a basic level. Practical, beyond-basic AI skills can therefore still help a job candidate stand out.

6. Contact center w Polsce. Jak wykorzystać pełnię potencjału dla klienta i organizacji? (Contact centres in Poland: how to use their full potential for customers and organisations)

EY-Parthenon with the Polish Direct Marketing Association (PSMB) · 16 September 2026 · survey of about 50 organisations and about 1,000 consumers

The study covers contact centres in banking, insurance, healthcare, energy and business-process outsourcing. AI is changing customer service through voicebots, advanced analytics and tools that support consultants in real time. 83% of organisations name cost reduction as the effect of introducing AI, but only 23% name better customer satisfaction. Only 17% of customers always prefer contact with a human.

For ReSTART: Customer-service agents in Poland will increasingly work alongside voicebots and AI assistants. The human role moves towards complex cases, quality checks and customer satisfaction, which are skills ReSTART can train.

Exposure of occupations to AI and the impact on the labour market

7. Generatywna sztuczna inteligencja a polski rynek pracy (Generative Artificial Intelligence and the Polish Labour Market) — research report and supplement

NASK – National Research Institute with the International Labour Organization (ILO), funded by the Ministry of Digital Affairs · June 2025 (newest edition) · task-based exposure index applied to Labour Force Survey data, worker survey (December 2024) and interviews

This is the most detailed Polish study of generative-AI exposure. It applies the ILO–NASK index (ILO Working Paper 140, May 2025, built with Polish task data and a survey of 1,640 workers in Poland) to Polish employment data and to 2,539 occupations of the Polish classification, sorting them into four exposure gradients (Gradient 4 = highest). The authors stress that exposure means that tasks can be automated or transformed — it does not mean that occupations will inevitably disappear.

About 30.3% of all jobs in Poland (about 5.08 million) show some exposure to generative AI, and 4.9% (817,500 workers) are in the highest-exposure gradient. Clerical (office) workers are the most exposed group: 71.2% work in exposed occupations and 47.2% are in the highest gradient — about 597,000 jobs. Exposure is much higher for women (39.1% of female employment) than for men (22.8%). The supplement places the call-centre worker, e-customer-service worker, secretary, office worker, receptionist, management assistant and social-media specialist in Gradient 3, and the telemarketer in Gradient 4. Globally, ILO finds that most of the 13 occupations in the highest gradient are clerical roles.

Actual use was still low at the end of 2024: only 9.4% of workers said GenAI tools had been introduced at their workplace, and 68.1% had received no guidelines on using them. The most common work use was writing and editing documents, content and e-mails (37.5%). 58.4% of working Poles say they are ready to learn about AI.

For ReSTART: This study is the core Polish evidence for ReSTART: all four ReSTART profiles are in exposed occupations, and clerical work is the most exposed group in Poland. Because exposure means task change, ReSTART should train people to do the remaining human tasks well — checking, deciding, handling exceptions and dealing with people.

8. OECD Employment Outlook 2026 — Figure 3.3: Regional exposure to generative AI

OECD · July 2026 (data file updated 7 July 2026; latest available year) · flagship OECD report, regional data

The OECD measures the share of employment that is highly exposed to generative AI now or in the near future, by region. For Poland, the national figure is 45.5%, ranging from 36.5% in Świętokrzyskie to 68.8% in the Warsaw region.

For ReSTART: AI exposure in Poland is concentrated in metropolitan labour markets with many office and service jobs. Outside the big cities, ReSTART learners may also find remote and hybrid work with employers based in these regions.

9. AI w polskich przedsiębiorstwach (AI in Polish enterprises), Point Paper 5/2025 — and AI na polskim rynku pracy (AI on the Polish labour market), 2024

Polish Economic Institute (PIE) · September 2025 and October 2024 (newest PIE publications on the topic) · policy papers combining statistics, surveys and interviews

PIE’s 2025 paper compares several surveys and finds that, depending on the sample and method, between 5.9% and 16% of Polish firms use AI; 77% of firms that do not use AI say they will not adopt it until it becomes necessary. AI does not yet have a large effect on employment: according to GUS data, only 5% of industrial firms using AI saw employment fall as a result. However, the paper reports signals that firms are changing what they expect from new hires; one interviewed employer said the company had largely given up junior and testing positions and now hires senior staff able to verify data produced by AI.

PIE’s 2024 study, the main earlier Polish estimate, found that about 3.68 million people (about 22% of workers covered) work in the 20 occupations most exposed to AI. Secretaries are among them, and administrative and specialised secretarial staff and customer-information clerks are highly exposed to large language models.

For ReSTART: Entry-level office tasks are the first to be handed to AI, so beginners must show from the start that they can supervise and verify AI work. This is an argument for including AI verification and quality control in every ReSTART profile.

10. Workers’ Exposure to AI Across Development Stages (IBS Working Paper 02/2025)

Institute for Structural Research (IBS), Warsaw · revised version October 2025 · econometric working paper based on OECD PIAAC data for 108 countries

The Warsaw-based institute builds a task-based measure of AI exposure from worker-level survey data and compares countries at different levels of development. Its key finding for ReSTART is that clerical support workers consistently show the highest AI exposure in all country income groups, and that exposure in high-income countries is rising because of changes in tasks within occupations.

For ReSTART: What matters is the task content of a job, not only the job title. ReSTART’s task-based lesson design (what to automate, what to check, what to keep human) fits this evidence.

11. Skills forecast 2025 — Poland

Cedefop (European Centre for the Development of Vocational Training) · 2025 (newest country report; forecast to 2035) · harmonised EU skills forecast

Cedefop forecasts that employment in Poland will grow by around 900,000 between 2022 and 2035 (+5.4%), and that there will be around 5.6 million job openings to replace people leaving the labour market. Despite this replacement demand, jobs are forecast to be lost in agriculture and, to a much lesser degree, among service and sales workers, clerks and elementary occupations. About two-thirds (64%) of all job openings to 2035 will require high-level qualifications.

For ReSTART: Clerical jobs will not disappear in Poland, but they will shrink slowly and qualification requirements will rise. Stackable micro-credentials help adults show higher-level skills without a full degree.

Demand for occupations and skills

12. Barometr zawodów 2026 (Occupational Barometer 2026) — national summary report and Lubelskie voivodeship report

Voivodeship Labour Office in Kraków (national coordinator) and in Lublin, commissioned by the Ministry of Family, Labour and Social Policy · December 2025 (newest edition; forecast for 2026) · expert-panel forecast in every county

The Barometer is the official annual forecast of shortage and surplus occupations, prepared in 345 expert panels with 2,482 experts across all counties (powiaty) and supported by an employer survey (3,963 responses). For 2026 it rates 17 occupations as in shortage, 151 as balanced and none as in surplus nationally.

All occupations closest to the ReSTART profiles are rated as balanced both nationally and in Lubelskie: administrative and office workers; secretaries and assistants; telephone and electronic customer-service workers; receptionists; PR, advertising, marketing and sales specialists; and online-sales workers. However, a surplus of administrative and office workers is expected in 47 counties (more than a year earlier), and the report explains surpluses mainly by many graduates of popular fields and, increasingly, by how easily AI tools can replace simple tasks. It also expects AI to reduce demand in jobs where AI and automation are cheaper, giving call-centre workers as an example.

In Lubelskie, 19 occupations are in shortage and 4 in surplus. The number of counties with a surplus of administrative and office workers rose from 4 to 7 (of 24), and of administration specialists from 7 to 10.

For ReSTART: ReSTART profiles are not shortage occupations in Poland: candidates compete for jobs, especially in office administration. AI skills, verified by micro-credentials, can be the difference that helps a candidate stand out.

13. Nowe czy stare zasady gry na rynku pracy. BKL 2025/26 (New or old rules of the game in the labour market — Human Capital Study 2025/26)

PARP and Jagiellonian University · 2026 (surveys conducted in 2025) · employer survey and adult population survey (n = 6,101)

The Human Capital Study (Bilans Kapitału Ludzkiego, BKL) is Poland’s largest regular survey of employers’ skill needs and adult learning; this is the first wave of its new edition, with further waves planned for 2027 and 2029.

13% of employers implement AI and 11% plan to. AI is used mostly for support processes in everyday office work, including writing (77% of AI-using employers). Employers expect that within three years 5% of current jobs will be fully and 22% partly replaced by machines, robots or AI; for office workers these shares are much higher — 7% and 40%. The main challenges in introducing AI concern people: 53% of organisations that consider AI important lack the required skills.

Among adults, only 9% use AI generators regularly, and only 7–8% use AI at work. At the same time, 79% of jobseekers say they are ready to retrain — more often women (84% vs 74% of men) and people with secondary (85%) or lower (81%) education than graduates (58%).

For ReSTART: Employers expect office work to change more than any other job group, and jobseekers — especially women and people with secondary education, ReSTART’s core learners — are very open to retraining. This is a strong case for ReSTART in Poland.

14. Bezrobotni zarejestrowani i stopa bezrobocia — stan w końcu sierpnia 2026 r.; Bezrobocie rejestrowane 1–2 kwartał 2026 r. (Registered unemployment, August 2026 and first half of 2026)

Statistics Poland (GUS) · 23–24 September 2026 · official register data from county labour offices

At the end of August 2026, 909,500 people were registered as unemployed in Poland, 6.2% more than a year earlier, and the registered unemployment rate was 5.8% (0.3 percentage points higher than a year earlier). In Lubelskie the rate was 8.1% (60,900 people). Data since June 2025 are collected under the new labour-market act, which affects comparisons over time.

At the end of June 2026, 49.8% of registered unemployed people were long-term unemployed, 26.6% were over 50, and 33.6% had no vocational qualifications; 87.5% had worked before.

For ReSTART: The typical registered unemployed person in Poland has work experience but often lacks a formal qualification, and half have been out of work for a long time. Short, recognised, practical training — such as ReSTART micro-credentials — matches this profile, especially in Lubelskie, where unemployment is above the national average.

15. Rynek Pracy Specjalistów. Podsumowanie I połowy 2026 (Specialist labour market: summary of the first half of 2026)

Pracuj.pl (Grupa Pracuj) · 20 July 2026 · portal data on job ads and applications (387,000 ads)

Poland’s largest specialist job portal reports on the 387,000 ads it published in January–June 2026. AI-related competences appeared in 4% of all ads (2% a year earlier) and in 6.3% of IT ads; the portal notes that employers look for people who can use AI tools not only in technology but also in marketing, sales, e-commerce and finance.

26% of ads offered hybrid work, 9% fully remote work and 65% on-site work. Customer service accounted for 7% of ads. Office administration attracted 15% of all applications (second place), which shows strong competition. Offers for beginners made up 14% of all ads; office administration (14%) and customer service (12%) were among the main entry areas. 32% of ads showed a salary range.

For ReSTART: AI-tool skills are a small but fast-growing requirement in Polish job ads. Customer service and office administration remain open doors for beginners, so ReSTART learners should name concrete AI skills in their CVs to stand out among many applicants.

16. Liczba ogłoszeń wymagających kompetencji AI wzrosła w Polsce o ok. 33 tys. w 2025 r. (Job ads requiring AI skills in Poland rose by about 33,000 in 2025) — AI Jobs Barometer 2026

PwC Polska (PwC 2026 Global AI Jobs Barometer) · 25 June 2026 · analysis of over one billion job ads in 27 countries

In Poland, job ads requiring AI skills grew by about 33,000 in 2025 and reached 3.5% of all ads (2% a year earlier). The growth is driven mainly by jobs that require knowledge of AI tools and their practical use, not only by AI developers. PwC distinguishes jobs that AI “professionalises” from jobs where AI lowers the entry barrier — the latter, which include administrative workers, see slower growth in job offers and pay. Globally, the average wage premium for AI skills reached 62% in 2026 (57% a year earlier).

For ReSTART: Administrative staff need to move from tasks AI makes easy for everyone towards tasks where AI is used with expertise — coordination, quality control and process improvement. This is the upskilling direction ReSTART should take.

17. Oferty pracy w Polsce. Edycja LXXIV: sierpień 2026 (Job offers in Poland, August 2026)

Grant Thornton Polska with Element · 15 September 2026 · monitoring of the 50 largest job portals and content analysis of 1,000 ads

In August 2026 employers published 257,046 new job ads on the 50 largest Polish job portals, 4% more than a year earlier. Training was offered as a benefit in 48% of ads, and specific professional experience was required in 59%.

For ReSTART: Because most employers ask for experience, ReSTART’s problem-based tasks and portfolio outputs give learners concrete proof of experience to show in applications.

18. Employed persons working from home (dataset lfsa_ehomp) — 2025 data

Eurostat (EU Labour Force Survey) · updated 10 September 2026 · official EU statistics

In 2025, 6.1% of employed Poles aged 15–64 usually worked from home and 10.2% sometimes did, compared with 8.8% and 14.2% in the EU. Remote work in Poland is less common than in the EU but is growing slowly (usually from home: 5.5% in 2024, 6.1% in 2025).

For ReSTART: In Poland, hybrid work is more common than fully remote work. ReSTART should prepare learners for hybrid arrangements and should not promise fully remote jobs, except for freelance virtual assistance.

Digital skills and AI literacy of adults

19. Digital Decade 2026 Country Report — Poland

European Commission, DG CONNECT (SWD(2026) 155 final, part 21/27) · 17 June 2026 · official EU monitoring report (data mostly for 2025)

The Commission’s annual assessment of Poland’s progress towards the EU Digital Decade 2030 targets, covering digital skills, infrastructure, digitalisation of businesses and public services. It uses harmonised Eurostat survey indicators (ICT usage by individuals and enterprises, reference year 2025), the national Digital Decade roadmap and the Special Eurobarometer on the Digital Decade 2026.

In 2025, 50.42% of Poles aged 16–74 had at least basic digital skills (44.3% in 2023), against an EU average of 60.4%, and the report notes that Poland is lagging behind its own national trajectory. The gap is widest among older adults: only 21.99% of people aged 55–74 have basic skills (EU 42.6%). Generative AI was used by 22.68% of people in Poland in 2025 (EU 32.66%), but only 8.36% used it for professional purposes (EU 15.36%). According to the Digital Decade Eurobarometer 2026, the main obstacle to using generative AI tools is a lack of training or relevant skills (29%), followed by not seeing a need (27%) and concerns about ethical use or misuse (24%). ICT specialists represent 4.5% of total employment (EU 5.0%).

On businesses, AI is used by 8.36% of Polish enterprises (EU 19.95%) and 7.09% of SMEs (EU 18.9%); the report states that Polish enterprises’ use of data analytics and AI is the lowest and second lowest in the EU respectively. The Commission recommends that Poland strengthen measures for social groups with lower digital skills, such as older adults and inhabitants of rural areas, and translate the revised national AI policy into concrete measures for key sectors.

For ReSTART: This is the most authoritative and recent benchmark for Poland. It shows a double gap that ReSTART addresses directly: basic digital skills are 10 percentage points below the EU average, and professional use of generative AI is about half the EU level. The finding that 29% of people point to missing training as the main barrier supports ReSTART’s short, practice-based AI modules.

20. Individuals’ level of digital skills (from 2021 onwards) — 2025 data for Poland, incl. unemployed people

Eurostat (dataset isoc_sk_dskl_i21, updated 17 April 2026) · 2025 reference year · official EU statistics (DigComp-based Digital Skills Indicator 2.0)

Eurostat’s harmonised survey on ICT usage in households and by individuals (people aged 16–74) measures digital skills across five DigComp areas: information and data literacy, communication and collaboration, digital content creation, safety, and problem solving. The indicator is collected every two years (2021, 2023, 2025) and can be broken down by labour-market status, age and education.

For ReSTART’s core target group the picture is stark: in 2025 only 39.9% of unemployed people in Poland had at least basic digital skills, compared with 54.42% of unemployed people across the EU (2023: 34.6% in Poland vs 47.71% in the EU). The share for all Polish adults was 50.42% (EU 60.4%), and for people aged 55–74 it was 21.99% (EU 42.6%).

The unemployed in Poland are therefore about 10 percentage points behind the Polish adult average and about 14.5 points behind unemployed people in the EU as a whole. The share is rising, but slowly.

For ReSTART: This is the clearest statistical justification for ReSTART in Poland: six in ten unemployed Poles lack basic digital skills, which comes before any AI skill. Learning paths should therefore start with core digital confidence (files, e-mail, online safety, simple data) and then build AI use on top.

21. Survey of Adults Skills 2023: Poland — Country Note (OECD PIAAC Cycle 2)

OECD, Directorate for Education and Skills · 10 December 2024 (newest edition of the survey; fieldwork 2022–23) · international assessment of adult skills

The second cycle of the OECD Survey of Adult Skills directly tested adults aged 16–65 on tablets in literacy, numeracy and adaptive problem solving (the ability to handle new, changing problems, which is close to working with digital and AI tools). This is the newest edition of the survey.

Adults in Poland scored 236 points in literacy, 239 in numeracy and 226 in adaptive problem solving, all below the OECD average. 39% of adults (OECD: 26%) scored at Level 1 or below in literacy, 38% (OECD: 25%) in numeracy and 48% (OECD: 29%) in adaptive problem solving; 27% (OECD: 18%) were at the two lowest levels in all three domains. Average results fell compared with 2011–12.

The OECD notes data-quality issues for a high share of Polish respondents and advises that results “should be interpreted with greater caution” than for other countries.

For ReSTART: Even allowing for the caveat, a large group of Polish adults has weak foundation skills. ReSTART materials for Poland should use plain language (A2/B1), short steps and worked examples, and should train the checking of AI outputs as a problem-solving skill, not only tool use.

22. Sztuczna inteligencja w Polsce. Krajobraz pełen paradoksów (Artificial intelligence in Poland: a landscape of paradoxes) — Polish results of Trust, Attitudes and Use of AI: A Global Study 2025

KPMG in Poland with the University of Melbourne · 2025 (fieldwork November 2024 – January 2025; newest edition) · online survey, 48,000 respondents in 47 countries, including 1,082 in Poland

Among working Poles, 54% use AI intentionally at least once every few months, and nearly 90% of users rely on publicly available tools. Only 29% have had any AI training — 10 percentage points below the global average — and 55% of workers who use AI do not reveal it, presenting AI output as their own.

For ReSTART: Much AI use in Polish workplaces is informal, untrained and hidden. ReSTART’s lessons on transparency, data protection and checking AI output respond directly to this risk.

23. Global AI Diffusion — Q2 2026 Trends and Insights

Microsoft AI Economy Institute · September 2026 · estimate based on anonymised device telemetry (not official statistics)

Microsoft estimates that 32.3% of Poland’s working-age population (15–64) used a generative AI product in Q2 2026, up from 31.0% in Q1, placing Poland 24th among the economies measured (global average 18.8%). The measure differs from the official surveys and is not directly comparable with them.

For ReSTART: This most recent signal suggests that everyday use of AI keeps spreading in Poland. Learners will increasingly meet colleagues and customers who use AI, which makes basic AI literacy part of employability.

Policy, strategies and public support

24. Polityka rozwoju sztucznej inteligencji w Polsce do 2030 roku (Policy for the Development of Artificial Intelligence in Poland until 2030) — draft of 15 April 2026

Ministry of Digital Affairs (Ministerstwo Cyfryzacji), draft resolution of the Council of Ministers (ID260) · published 16 April 2026 · national strategic policy (draft)

The draft policy is meant to replace the AI policy adopted by the Council of Ministers in January 2021. The current draft, dated 15 April 2026, went through inter-ministerial agreement, opinions and consultation in April–May 2026; the government’s status page (as of 30 June 2026) planned an update of the text in August 2026, and the European Commission expects adoption in the second half of 2026. We found no record of adoption by 1 October 2026.

Goal 2 of the draft is a society with above-basic competences in the practical use of AI: at least 30% of Polish society with above-basic practical AI skills, and at least 10 programmes adapting the labour market to AI (reskilling and upskilling). Goal 4 targets at least 34% of Polish SMEs using AI tools in their business processes. The labour-market chapter states that support will include mapping trends, creating jobs in new occupations and helping people gain qualifications that keep them active in work (upskilling, reskilling).

The draft’s monitoring chapter cites the NASK–ILO finding that about 30% of all jobs in Poland are already exposed to generative AI, notes that office (clerical) occupations are the most exposed (citing ILO 2023: 24% of tasks highly and a further 58% moderately exposed) and expects falling demand for routine work, naming customer-service workers as an example. It also cites a forecast that the Polish labour force will shrink by 2.1 million workers (12.6% of current employment) by 2035.

For ReSTART: The draft policy names exactly the occupations ReSTART targets (office and customer-service work) as most affected by AI and sets national targets for AI skills and reskilling programmes. ReSTART can present itself as a ready-made, tested reskilling pathway that matches Goal 2.

25. Ustawa z dnia 3 lipca 2026 r. o systemach sztucznej inteligencji (Act of 3 July 2026 on Artificial Intelligence Systems)

Sejm of the Republic of Poland; drafted by the Ministry of Digital Affairs · Journal of Laws (Dz.U.) 2026 item 1003, published 27 July 2026 · legal act

The Act implements the EU AI Act (Regulation 2024/1689) in Polish law. It sets up the Commission for the Development and Security of Artificial Intelligence (Komisja Rozwoju i Bezpieczeństwa Sztucznej Inteligencji, KRiBSI) as the national market-surveillance authority for AI systems, regulates procedures and penalties, and includes measures to support AI development.

Among the Commission’s tasks the Act lists publications, educational programmes popularising knowledge about AI, and information activities. The Act enters into force 14 days after publication, with several chapters applying three months after publication.

For ReSTART: From 2026 Polish employers work under a national AI-Act framework. Basic knowledge of responsible AI use (data protection, transparency, human oversight) is now a practical workplace expectation, which ReSTART’s governance and verification lessons cover.

26. PLLuM ułatwi życie (m)obywatelom (PLLuM will make life easier for citizens) — Polish language model in the mObywatel app

NASK – National Research Institute · 30 December 2025 · official news release

PLLuM is a family of open Polish large language models developed by a state-funded consortium; the models and portal were made available in 2024, and a follow-up project (HIVE2) was contracted on 21 May 2026 according to the government priority page. From 31 December 2025, a PLLuM-based virtual assistant (chatbot) answers citizens’ questions in the mObywatel public-services app.

For ReSTART: Many Polish adults will meet AI first through public services. ReSTART can use this familiar example to explain what a chatbot can and cannot do, and why answers must be checked.

27. Ustawa z dnia 20 marca 2025 r. o rynku pracy i służbach zatrudnienia (Act of 20 March 2025 on the Labour Market and Employment Services)

Sejm of the Republic of Poland / Ministry of Family, Labour and Social Policy · Journal of Laws 2025 item 620, published 12 May 2025, in force from 1 June 2025 · legal act

The Act governs public employment services, including county labour offices (powiatowe urzędy pracy), and the support they offer. Its training instruments for unemployed people and jobseekers, financed from the Labour Fund, include self-chosen or commissioned training lasting up to 24 months, a continuing-education voucher (bon na kształcenie ustawiczne – the older term “training voucher” no longer appears in the Act), an interest-free education loan of up to 400% of the average wage, and funding for validation and certification of skills.

Since 1 January 2026, training financed in this way must be delivered by providers registered in PARP’s Development Services Database (Baza Usług Rozwojowych, BUR). The National Training Fund (KFS) is a separate instrument that supports continuing education of people in work.

For ReSTART: For a learner in Poland, the county labour office is the gateway to funded training. ReSTART courses and micro-credentials will be easier to use in Poland if training providers delivering them are registered in BUR and if learners know how to ask for a training voucher.

28. Priorytety wydatkowania Krajowego Funduszu Szkoleniowego na 2026 rok (National Training Fund spending priorities for 2026)

Priorities set by the minister responsible for labour; published by county labour offices on the official praca.gov.pl portal (e.g., PUP Gdynia, 9 January 2026) · 2026 · public funding priorities

The National Training Fund (Krajowy Fundusz Szkoleniowy, KFS) co-finances continuing education of employees, employers, self-employed people and contractors, with applications submitted through county labour offices. One of the four national priorities for 2026 is support for continuing education linked to new processes, technologies and work tools in firms, with particular attention to digital skills, AI and green skills, especially where these changes threaten job loss.

For ReSTART: Employers in Poland can co-finance AI upskilling of staff at risk of job loss through KFS in 2026. This is a sustainability route for ReSTART modules after the project, for example for administrative and customer-service staff.

29. Kluby Rozwoju Cyfrowego (Digital Development Clubs) — scaling phase and the “Buduj cyfrową formę – klik po kliku” campaign

Digital Poland Projects Centre (CPPC) / Ministry of Digital Affairs · 18 March 2026 (call) and 7 May 2026 (campaign) · public programmes (EU funds FERS; Recovery and Resilience Plan)

Digital Development Clubs are local places in municipalities where educators help adults find their way in the digital world. The scaling call opened on 18 March 2026 has total funding of over PLN 975.8 million and plans 2,066 new clubs; according to the European Commission, the aim is to train 450,000 adults by the end of 2028. Participants can also get an explanation of basic uses of AI.

In parallel, the Ministry’s national campaign “Buduj cyfrową formę – klik po kliku” (Build your digital fitness, click by click), financed from the Recovery and Resilience Plan, includes a strand “Odśwież cyfrowe umiejętności” (Refresh your digital skills) aimed at people returning to the labour market, with self-check tools at cyfrowaforma.gov.pl.

For ReSTART: These programmes reach adults with low digital skills locally but treat AI only at a basic level. ReSTART can complement them: a learner who has completed a club or the self-check can continue with ReSTART’s job-specific AI pathway.

The four ReSTART job profiles in Poland

This section summarises evidence for the four ReSTART profiles. Salary figures are gross monthly pay in PLN. Official GUS data (October 2024, published February 2026) are the most reliable but are available only for 3-digit occupation groups. The Sedlak & Sedlak national salary survey (updated January 2026) is a large online survey, not a random sample, and reports total pay including bonuses. Hays figures come from its own recruitment processes and reflect mostly larger employers. For context: the minimum wage is PLN 4,806 per month from 1 January 2026 (PLN 31.40 per hour for civil-law service contracts) [34], and the average gross wage in the enterprise sector was PLN 9,259.54 in August 2026 [35].

Exposure gradients come from the NASK–ILO study [7] (Gradient 1 = low, Gradient 4 = highest exposure to generative AI). Job-offer counts are keyword searches on Pracuj.pl, the largest Polish specialist job board; they include related job titles and change daily.

Virtual Assistant

Polish terms: wirtualna asystentka / wirtualny asystent, asystent zdalny · ReSTART course lead: POLARIS (IT)

For ReSTART: Present the virtual-assistant pathway in Poland as freelancing: include basics of self-employment, pricing, contracts and client data protection, and do not promise a salaried market rate.

Customer Service Agent

Polish terms: konsultant/ka obsługi klienta, specjalista/ka ds. obsługi klienta, konsultant/ka call center (contact center) · ReSTART course lead: HST (BG)

For ReSTART: Train agents to work with AI assistants: handle escalated and complex cases, check AI-suggested answers, write clear summaries and protect customer data.

Administrative Support Specialist

Polish terms: pracownik administracyjno-biurowy, asystent/ka biura, office assistant, referent, specjalista ds. administracyjnych · ReSTART course lead: WSEI (PL)

For ReSTART: This is the most AI-exposed and most competitive profile. ReSTART should move learners towards tasks that AI does not replace: coordinating work, checking documents and data, handling exceptions and improving office processes with AI.

Social Media Assistant

Polish terms: asystent/ka ds. social media, (młodszy) specjalista ds. social media / mediów społecznościowych · ReSTART course lead: EureCons (DE)

For ReSTART: Pure social-media jobs at assistant level are few, and the tasks are often part of broader marketing roles. ReSTART should combine AI-assisted content skills with basic marketing, analytics and copyright awareness.

Salary sources: [30] GUS, Structure of wages and salaries by occupation, October 2024; [31] Sedlak & Sedlak, national salary survey (OBW), position pages updated January 2026; [32] Hays Poland Salary Report 2026; [33] Pracuj.pl search results, 1 October 2026, 14:32 CEST; [34] Ministry of Family, Labour and Social Policy, minimum wage 2026; [35] GUS, average wage in the enterprise sector, August 2026.

Key institutions and stakeholders in Poland

General institutional contacts only (head-office address and general e-mail or phone), as published on the institutions’ official websites on 1 October 2026. No personal contact data are included.

Why this matters for ReSTART

The Polish evidence points in one direction: AI is changing the tasks of exactly the jobs ReSTART prepares people for, while the adults who most need these jobs have the weakest digital and AI skills. The main conclusions for the project and for the AI Job Market Trends booklet are:

• Transformation, not disappearance: Office and customer-service occupations are the most exposed to generative AI in Poland [7], and employers expect 40% of office jobs to be partly replaced within three years [13]. Yet fewer than 1% of firms have cut jobs because of AI [3], and all four profiles remain balanced occupations [12]. The booklet should say clearly: AI changes these jobs; it does not simply end them.

• The skills gap of the target group: Only 39.9% of unemployed Poles have basic digital skills [20], and unemployed and secondary-educated adults rarely use AI tools [1]. ReSTART pathways in Poland should start with basic digital confidence and plain-language AI literacy.

• Competition and the value of proof: Office administration attracts many applicants [15] and faces local surpluses, including in Lubelskie [12]. AI-tool skills appear in a small but fast-growing share of job ads [15][16]. Micro-credentials and portfolio outputs from problem-based tasks give learners evidence that employers can see.

• Responsible use: Most Polish workers who use AI have had no training, and many hide their use [22]; most employees received no workplace guidance [7]. Verification, data protection and transparency should be visible learning outcomes in every profile.

• Realistic work arrangements: Hybrid work is more common than fully remote work [15][18]. Salaried virtual-assistant jobs barely exist, so the VA pathway should be presented as freelancing.

• Policy alignment and sustainability: The draft national AI policy targets above-basic AI skills for at least 30% of society and at least 10 reskilling programmes [24]; the 2026 KFS priorities fund AI and digital upskilling of people at risk of job loss [28]; and the labour-market act lets county labour offices fund training and vouchers for unemployed people [27]. ReSTART results can be offered through these channels after the project ends.

Sources

[1] Statistics Poland (GUS), Statistical Office in Szczecin. Społeczeństwo informacyjne w Polsce w 2025 r. (Information society in Poland in 2025). 16 December 2025 (first release 21 October 2025; tables 8 December 2025). https://stat.gov.pl/files/gfx/portalinformacyjny/pl/defaultaktualnosci/5497/1/19/1/spoleczenstwo_informacyjne_w_polsce_2025.pdf

[2] Eurostat. 20% of EU enterprises use AI technologies — and The use of artificial intelligence (AI) technologies in the European Union: Key results, 2026 edition. 11 December 2025 (news) and spring 2026 (statistical report). https://ec.europa.eu/eurostat/web/products-eurostat-news/w/ddn-20251211-2

[3] Narodowy Bank Polski (National Bank of Poland). Szybki Monitoring NBP. Analiza sytuacji sektora przedsiębiorstw, nr 03/26 (NBP Quick Monitoring Survey, July 2026). 20 July 2026. https://nbp.pl/wp-content/uploads/2026/07/Szybki-Monitoring-NBP-lipiec-2026.pdf

[4] Polish Agency for Enterprise Development (PARP) and Jagiellonian University, Human Capital Study (BKL). Sztuczna inteligencja w firmach: gotowość do adopcji, kompetencje i potrzeby. Raport z badań BKL (Artificial intelligence in companies: readiness, competences and needs). 2026 (survey June–September 2025). https://www.parp.gov.pl/storage/publications/pdf/Badanie-AI—raport-podsumowujacy_21_04_2026.pdf

[5] Pracuj.pl (Grupa Pracuj), fieldwork by ARC Rynek i Opinia. Badanie Pracuj.pl: 70% pracowników nie korzysta z AI w pracy (Pracuj.pl survey: 70% of workers do not use AI at work) — “To tylko AI” study. 16 September 2026. https://media.pracuj.pl/471416-badanie-pracujpl-70-pracownikow-nie-korzysta-z-ai-w-pracy-realia-kontra-wizja-powszechnej-automatyzacji

[6] EY-Parthenon with the Polish Direct Marketing Association (PSMB). Contact center w Polsce. Jak wykorzystać pełnię potencjału dla klienta i organizacji? (Contact centres in Poland: how to use their full potential for customers and organisations). 16 September 2026. https://www.ey.com/pl_pl/insights/raporty-analizy/contact-center-w-polsce-gc-fy-27

[7] NASK – National Research Institute with the International Labour Organization (ILO), funded by the Ministry of Digital Affairs. Generatywna sztuczna inteligencja a polski rynek pracy (Generative Artificial Intelligence and the Polish Labour Market) — research report and supplement. June 2025 (newest edition). https://www.nask.pl/media/2025/06/Generatywna-sztuczna-inteligencja-a-polski-rynek-pracy.pdf

[8] OECD. OECD Employment Outlook 2026 — Figure 3.3: Regional exposure to generative AI. July 2026 (data file updated 7 July 2026; latest available year). https://www.oecd.org/en/publications/oecd-employment-outlook-2026_7e710f54-en.html

[9] Polish Economic Institute (PIE). AI w polskich przedsiębiorstwach (AI in Polish enterprises), Point Paper 5/2025 — and AI na polskim rynku pracy (AI on the Polish labour market), 2024. September 2025 and October 2024 (newest PIE publications on the topic). https://pie.net.pl/wp-content/uploads/2025/09/PointP-5-2025_AI-w-Firmach.pdf

[10] Institute for Structural Research (IBS), Warsaw. Workers’ Exposure to AI Across Development Stages (IBS Working Paper 02/2025). revised version October 2025. https://ibs.org.pl/wp-content/uploads/2025/03/Workers_AI_exposure_across_development_IBS_WP_202502.pdf

[11] Cedefop (European Centre for the Development of Vocational Training). Skills forecast 2025 — Poland. 2025 (newest country report; forecast to 2035). https://www.cedefop.europa.eu/files/skills_forecast_-_poland_2025.pdf

[12] Voivodeship Labour Office in Kraków (national coordinator) and in Lublin, commissioned by the Ministry of Family, Labour and Social Policy. Barometr zawodów 2026 (Occupational Barometer 2026) — national summary report and Lubelskie voivodeship report. December 2025 (newest edition; forecast for 2026). https://barometrzawodow.pl/forecast-card-zip/2026/report_pl/raport_ogolnopolski_2026.pdf

[13] PARP and Jagiellonian University. Nowe czy stare zasady gry na rynku pracy. BKL 2025/26 (New or old rules of the game in the labour market — Human Capital Study 2025/26). 2026 (surveys conducted in 2025). https://www.parp.gov.pl/component/publications/publication/nowe-czy-stare-zasady-gry-na-rynku-pracy-bkl-2025-26

[14] Statistics Poland (GUS). Bezrobotni zarejestrowani i stopa bezrobocia — stan w końcu sierpnia 2026 r.; Bezrobocie rejestrowane 1–2 kwartał 2026 r. (Registered unemployment, August 2026 and first half of 2026). 23–24 September 2026. https://ssgk.stat.gov.pl/Rynek_pracy.html

[15] Pracuj.pl (Grupa Pracuj). Rynek Pracy Specjalistów. Podsumowanie I połowy 2026 (Specialist labour market: summary of the first half of 2026). 20 July 2026. https://media.pracuj.pl/465266-rynek-pracy-specjalistow-w-i-polowie-2026-roku-rosnaca-rola-ai-i-nowe-standardy-jawnosci-raport-od-pracujpl

[16] PwC Polska (PwC 2026 Global AI Jobs Barometer). Liczba ogłoszeń wymagających kompetencji AI wzrosła w Polsce o ok. 33 tys. w 2025 r. (Job ads requiring AI skills in Poland rose by about 33,000 in 2025) — AI Jobs Barometer 2026. 25 June 2026. https://www.pwc.pl/pl/media/2026/2026-06-25-liczba-ogloszen-wymagajacych-kompetencji-ai-wzrosla-w-polsce-o-okolo-33-tys-w-2025-roku.html

[17] Grant Thornton Polska with Element. Oferty pracy w Polsce. Edycja LXXIV: sierpień 2026 (Job offers in Poland, August 2026). 15 September 2026. https://grantthornton.pl/wp-content/uploads/2026/09/OFERTY-PRACY-sierpien-2026.pdf

[18] Eurostat (EU Labour Force Survey). Employed persons working from home (dataset lfsa_ehomp) — 2025 data. updated 10 September 2026. https://ec.europa.eu/eurostat/databrowser/view/lfsa_ehomp/default/table?lang=en

[19] European Commission, DG CONNECT (SWD(2026) 155 final, part 21/27). Digital Decade 2026 Country Report — Poland. 17 June 2026. https://digital-strategy.ec.europa.eu/en/factpages/polands-2026-digital-decade-country-report

[20] Eurostat (dataset isoc_sk_dskl_i21, updated 17 April 2026). Individuals’ level of digital skills (from 2021 onwards) — 2025 data for Poland, incl. unemployed people. 2025 reference year. https://ec.europa.eu/eurostat/databrowser/view/isoc_sk_dskl_i21/default/table?lang=en

[21] OECD, Directorate for Education and Skills. Survey of Adults Skills 2023: Poland — Country Note (OECD PIAAC Cycle 2). 10 December 2024 (newest edition of the survey; fieldwork 2022–23). https://www.oecd.org/en/publications/survey-of-adults-skills-2023-country-notes_ab4f6b8c-en/poland_9d76682b-en.html

[22] KPMG in Poland with the University of Melbourne. Sztuczna inteligencja w Polsce. Krajobraz pełen paradoksów (Artificial intelligence in Poland: a landscape of paradoxes) — Polish results of Trust, Attitudes and Use of AI: A Global Study 2025. 2025 (fieldwork November 2024 – January 2025; newest edition). https://kpmg.com/pl/pl/wiedza/transformacja-biznesowa/zaufanie-do-sztucznej-inteligencji-w-polsce.html

[23] Microsoft AI Economy Institute. Global AI Diffusion — Q2 2026 Trends and Insights. September 2026. https://www.microsoft.com/en-us/research/wp-content/uploads/2026/09/Microsoft-AI-Diffusion-Report-2026-Q2.pdf

[24] Ministry of Digital Affairs (Ministerstwo Cyfryzacji), draft resolution of the Council of Ministers (ID260). Polityka rozwoju sztucznej inteligencji w Polsce do 2030 roku (Policy for the Development of Artificial Intelligence in Poland until 2030) — draft of 15 April 2026. published 16 April 2026. https://www.gov.pl/web/cyfryzacja/projekt-uchwaly-rady-ministrow-w-sprawie-ustanowienia-polityki-rozwoju-sztucznej-inteligencji-w-polsce-do-2030-roku-id260

[25] Sejm of the Republic of Poland; drafted by the Ministry of Digital Affairs. Ustawa z dnia 3 lipca 2026 r. o systemach sztucznej inteligencji (Act of 3 July 2026 on Artificial Intelligence Systems). Journal of Laws (Dz.U.) 2026 item 1003, published 27 July 2026. https://eli.gov.pl/eli/DU/2026/1003/ogl

[26] NASK – National Research Institute. PLLuM ułatwi życie (m)obywatelom (PLLuM will make life easier for citizens) — Polish language model in the mObywatel app. 30 December 2025. https://www.nask.pl/aktualnosci/pllum-ulatwi-zycie-mobywatelom

[27] Sejm of the Republic of Poland / Ministry of Family, Labour and Social Policy. Ustawa z dnia 20 marca 2025 r. o rynku pracy i służbach zatrudnienia (Act of 20 March 2025 on the Labour Market and Employment Services). Journal of Laws 2025 item 620, published 12 May 2025, in force from 1 June 2025. https://eli.gov.pl/eli/DU/2025/620/ogl

[28] Priorities set by the minister responsible for labour; published by county labour offices on the official praca.gov.pl portal (e.g., PUP Gdynia, 9 January 2026). Priorytety wydatkowania Krajowego Funduszu Szkoleniowego na 2026 rok (National Training Fund spending priorities for 2026). 2026. https://gdynia.praca.gov.pl/-/priorytety-wydatkowania-krajowego-funduszu-szkoleniowego-na-2026-rok

[29] Digital Poland Projects Centre (CPPC) / Ministry of Digital Affairs. Kluby Rozwoju Cyfrowego (Digital Development Clubs) — scaling phase and the “Buduj cyfrową formę – klik po kliku” campaign. 18 March 2026 (call) and 7 May 2026 (campaign). https://www.gov.pl/web/cppc/uruchomienie-naboru-FERS0109IP0300126

[30] Statistics Poland (GUS). Struktura wynagrodzeń według zawodów za październik 2024 r. (Structure of wages and salaries by occupation, October 2024). Published 10 February 2026. https://stat.gov.pl/obszary-tematyczne/rynek-pracy/pracujacy-zatrudnieni-wynagrodzenia-koszty-pracy/struktura-wynagrodzen-wedlug-zawodow-za-pazdziernik-2024-r-,4,12.html

[31] Sedlak & Sedlak. Ogólnopolskie Badanie Wynagrodzeń (national salary survey), position pages, last updated January 2026 (e.g., customer service specialist, call-centre consultant, administrative worker, secretary, social media manager, marketing assistant). https://wynagrodzenia.pl/moja-placa/ile-zarabia-specjalista-ds-obslugi-klienta

[32] Hays Poland. Raport płacowy 2026 (Salary Report 2026), based on Hays recruitments in 2025. https://www.hays.pl/raport-placowy/o-raporcie

[33] Pracuj.pl. Job search results for “konsultant obsługi klienta”, “pracownik administracyjno-biurowy”, “asystentka biura”, “social media”, “junior social media” and “wirtualna asystentka”, with remote and hybrid filters. https://www.pracuj.pl/praca

[34] Ministry of Family, Labour and Social Policy. Minimalne wynagrodzenie za pracę / Minimalna stawka godzinowa (minimum wage and minimum hourly rate 2026). https://www.gov.pl/web/rodzina/minimalne-wynagrodzenie-za-prace

[35] Statistics Poland (GUS). Przeciętne zatrudnienie i wynagrodzenie w sektorze przedsiębiorstw w sierpniu 2026 r. (Average employment and wages in the enterprise sector, August 2026). 21 September 2026. https://stat.gov.pl/download/gfx/portalinformacyjny/pl/defaultaktualnosci/5474/3/177/1/przecietne_zatrudnienie_i_wynagrodzenie_w_sektorze_przedsiebiorstw_w_sierpniu_2026_r..pdf