The WIG20 Scoring Sheet
The WIG20 Scoring Sheet
The evidence behind “Not One WIG20 Supervisory Board Says It Oversees AI.” For each of the twenty companies: what its reports say about AI, with the page and the passage.
The essay Not One WIG20 Supervisory Board Says It Oversees AI. reads two documents from each WIG20 company: its annual report for 2025 and, where it has one, its supervisory board’s own report on 2025. This page shows what each says about AI, and where.
A score shows what a report discloses, not what the company does.
- 6name who decides about AI
- 3say a board oversees AI
- 1state a rule that can stop AI
- 6mention AI
- 3took note of information on AI
- 0say the board oversees AI
All twenty, A to Z
| Company | Annual report decides · oversees · stop | Supervisory board’s own report |
|---|---|---|
| Alior Bank | Mentions AI | |
| Allegro.eu | No supervisory board | |
| Bank Pekao | Took note of AI | |
| Budimex | No mention of AI | |
| CD PROJEKT | No mention of AI | |
| Dino Polska | No mention of AI | |
| Erste Bank Polskaex-Santander | No mention of AI | |
| Grupa Kęty | No mention of AI | |
| KGHM | No mention of AI | |
| KRUK | Mentions AI | |
| LPP | No mention of AI | |
| mBank | Mentions AI | |
| MODIVOex-CCC | No mention of AI | |
| ORLEN | Took note of AI | |
| Pepco Group | No supervisory board | |
| PGE | No mention of AI | |
| PKO BP | Took note of AI | |
| PZU | No mention of AI | |
| TAURON | No mention of AI | |
| Żabka | No supervisory board |
Each company’s section gives the verdict, the scores and both reports; the passages, in the language of the report and in translation, open under the toggle. How the reports were read and scored is at the end.
Alior Bank
No owner of AI is named. The closest is a model risk committee, in a passage that ties model risk to machine learning. No board is said to oversee AI, and no rule can stop an AI tool. The supervisory board's report mentions AI only as staff training.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: Alior Bank S.A.
Read: Sprawozdanie Zarządu z działalności Grupy Kapitałowej Alior Banku w 2025 roku obejmujące Sprawozdanie Zarządu z działalności Alior Banku oraz Oświadczenie dotyczące zrównoważonego rozwoju, PDF, 306 pages, published 24 February 2026, in the package aliorbank.pl, linked from aliorbank.pl. The consolidated financial statements and the supervisory board's assessment in the same package do not mention AI.
Who decides about AI?PartialPage 88, risk report, on model risk.
“Zarówno metodyki budowy, monitoringu jak i walidacji są stale rozwijane i doskonalone w odpowiedzi na zmieniające się otoczenie biznesowe i technologiczne, w szczególności ze względu na stosowanie technik Machine Learning w modelowaniu ryzyka. Ryzyko modeli nadzorowane jest operacyjnie przez Komitet Ryzyka Modeli.”
The methods for building, monitoring and validating models are continuously developed and improved in response to the changing business and technological environment, in particular because Machine Learning techniques are used in risk modelling. Model risk is overseen operationally by the Model Risk Committee.
A model risk committee is named, and the same passage ties model risk to machine learning. The report names no owner of AI.
Does the board oversee AI?Not foundThe nearest passage: the supervisory board's Risk Committee received periodic risk reports that include model risk (page 119). That passage does not mention AI or machine learning. The link to machine learning sits in the risk report (page 88), a separate passage, so it is not scored.
“Komitet otrzymywał okresowe raporty z obszaru ryzyka, zawierające informacje dotyczące ryzyka kredytowego, adekwatności kapitałowej, ryzyka rynkowego, ryzyka płynności, ryzyka operacyjnego, ryzyka technologicznego, ryzyka modeli, a także ryzyka braku zgodności i ryzyka reputacji”
The Committee received periodic risk reports with information on credit risk, capital adequacy, market risk, liquidity risk, operational risk, technology risk and model risk, as well as compliance risk and reputational risk
Is there a rule that can stop AI?Not foundThe nearest passage is general model validation (page 88), with no gate specific to AI and no stop outcome.
“modele rozpoznawane jako istotne dodatkowo są cyklicznie walidowane przez niezależną jednostkę”
models recognised as material are also validated at regular intervals by an independent unit
The supervisory board’s own report on 2025
Sprawozdanie z działalności Rady Nadzorczej Alior Banku S.A. w 2025 roku, PDF, 62 pages, a scan read with optical character recognition, attachment 6 to the notice of the general meeting of 29 April 2026, bonnier.pl. Page 25, in the account of the bank's spending on education, among staff programmes:
“ASIA to program rozwojowy skierowany do wybranych pracowników zainteresowanych pogłębieniem kompetencji w obszarze sztucznej inteligencji.”
ASIA is a development programme for selected employees who want to deepen their skills in artificial intelligence.
The same page lists AI webinars, a Copilot competition and AI meet-ups for staff. Training, not oversight.
Allegro.eu
Policies, not owners. The Board of Directors approves the AI Principles along with all other policies, approved a risk analysis that lists an AI risk, and oversees corporate risk in general. AI Ambassadors advise staff. The AI Principles cover risk and prohibited practices, but no rule to stop an AI system.
Read Annual report 2025
Show the evidence, with pages
Company: Allegro.eu S.A.
Read: Annual Report of Allegro.eu S.A. Group for the year ended 31 December 2025, PDF reading version, March 2026, 256 PDF pages laid out as two-page spreads (printed pages 1-512), about.allegro.eu, listed at about.allegro.eu. The PDF states that the binding version is the ESEF package, which was not searched. The financial statements in the same PDF were searched too. Pages below give the printed spread, then the PDF page.
Who decides about AI?PartialPages 332-333 (PDF page 167) and 336-337 (PDF page 169), sustainability statement.
The list of business conduct policies includes “Artificial Intelligence Principles” and states:
“All above policies are reviewed and approved by the Board of Directors at least once a year or once in two years and are available to all employees.”
“Furthermore, Allegro has appointed AI Ambassadors – designated points of contact for all employees regarding the use of AI within the company. These ambassadors provide guidance, answer questions, and help in AI adoption.”
The Board of Directors approves all listed policies, the AI Principles among them. The AI Ambassadors are advisory. The report names no owner of decisions on AI.
Does the board oversee AI?PartialPages 330-331 (PDF page 166) and 126-127 (PDF page 64).
“The Board of Directors approved the IRO analysis, which was the final step in the process.”
The table of material impacts, risks and opportunities on the same spread lists the risk “Insufficient corporate governance and ethical frameworks on artificial intelligence (AI) development”. In the risk framework, the Board of Directors' role includes “Oversight of corporate risk”.
The Board approved a risk analysis that lists an AI risk, and it oversees corporate risk in general. The report does not say that the Board or a board committee reviewed AI risk as such.
Is there a rule that can stop AI?PartialPages 336-337 (PDF page 169), on the AI Principles.
“including a risk-based approach in the management of AI systems, specific AI practices that are prohibited due to their potential to violate fundamental human rights, as well as regulatory requirements on the high-risk AI systems.”
“It also includes procedures for reporting new AI implementations, guidelines for the use of generative AI tools, principles of transparency and labeling of AI-generated content”
The report states no rule for stopping or withdrawing an AI system.
The supervisory board’s own report on 2025
Allegro.eu has a single board of directors and no supervisory board. Its board committees report inside the annual report above.
Bank Pekao
The closest supervisory board. It took note of the AI Policy that the management board adopted as part of the risk management system, and so did its Risk Committee. The annual report names an AI Transformation Department without a remit, ties AI to a code of conduct the management board checks, and plans rules for assessing AI risk. Neither document says that a board oversees AI or names a rule to stop it.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: Bank Polska Kasa Opieki S.A. (Bank Pekao).
Read: Sprawozdanie z działalności Grupy Kapitałowej Banku Pekao S.A. za 2025 rok, PDF, 347 pages, published 19 February 2026, pekao.com.pl. The consolidated financial statements do not mention AI.
Who decides about AI?PartialPages 14 and 95.
“Pion Technologii i Operacji został przekształcony w Pion Transformacji Technologicznej i Innowacji, w skład którego wchodzą Centrum IT, Centrum Bezpieczeństwa Banku oraz Departament Nowych Technologii i Rozwoju pod nową nazwą Departament Transformacji AI.”
The Technology and Operations Division was turned into the Technology Transformation and Innovation Division, made up of the IT Centre, the Bank Security Centre and the New Technologies and Development Department under its new name, the AI Transformation Department.
“Wiceprezes Zarządu Banku Marcin Zygmanowski nadzoruje działalność Pionu Transformacji Technologicznej i Innowacji, jak również koordynuje działania mające na celu odpowiednie zarządzanie ryzykiem dotyczącym bezpieczeństwa środowiska teleinformatycznego.”
Vice-President of the Management Board Marcin Zygmanowski oversees the Technology Transformation and Innovation Division and coordinates the work of managing the risk to the security of the ICT environment.
A department carries AI in its name. The report does not describe its remit. A management board member oversees the division the department belongs to.
Does the board oversee AI?PartialPage 335, with page 85.
Page 335, the section on the Code of Conduct:
“budowania godnej zaufania i etycznej sztucznej inteligencji poprzez stosowanie wytycznych dotyczących transparentności i niedyskryminacji”
building trustworthy and ethical artificial intelligence by applying guidelines on transparency and non-discrimination
“Zarząd na podstawie informacji przekazywanej przez Departament Zgodności okresowo weryfikuje i ocenia przestrzeganie zasad etyki określonych w Kodeksie Postępowania, w celu dostosowania ich do zmieniającej się sytuacji wewnętrznej Banku i otoczenia. Informacja o wynikach tej oceny przekazywana jest Radzie Nadzorczej.”
On the basis of information from the Compliance Department, the Management Board periodically checks and assesses compliance with the ethical principles set out in the Code of Conduct, in order to adapt them to the changing situation inside and around the Bank. Information on the results of this assessment is passed to the Supervisory Board.
Page 85, the governance statement lists among the Code's topics:
“etyczne wykorzystanie technologii i innowacji, w tym AI”
ethical use of technology and innovation, including AI
The management board checks compliance with a code of conduct that covers AI and reports the results to the supervisory board. The report does not say that either board oversees AI as such.
Is there a rule that can stop AI?PartialPage 321, the bank's ICT security strategy for 2025-2027.
“Sztuczna inteligencja i uczenie maszynowe (AI/ML) - wdrażanie mechanizmów bezpieczeństwa oraz zasad oceny ryzyka dla nowych technologii.”
Artificial intelligence and machine learning (AI/ML): introducing security mechanisms and risk assessment rules for new technologies.
“Prowadzimy analizę ryzyka oraz opracowujemy zasady jego oceny, które będą stosowane w procesach wdrożeniowych.”
We are analysing risk and developing rules for assessing it, which will be applied in deployment processes.
Both are plans. The report gives no criterion for stopping an AI tool.
The supervisory board’s own report on 2025
Sprawozdanie Rady Nadzorczej Banku for 2025, PDF, 60 pages, in the materials for the general meeting of 28 May 2026, pekao.com.pl. Page 10, the board's own work:
“Rada Nadzorcza zapoznała się z informacją przyjęcia przez Zarząd Banku Pekao S.A. „Polityki AI w Banku Pekao S.A.” jako elementu systemu zarządzania ryzykiem.”
The Supervisory Board took note of the information that the Management Board of Bank Pekao S.A. had adopted the "AI Policy at Bank Pekao S.A." as an element of the risk management system.
Page 44, in the Risk Committee's list of matters for 2025 (pages 42-45): “przyjęcie do wiadomości Polityki AI w Banku Pekao S.A.” (taking note of the AI Policy). Page 29, in the Audit Committee's list (pages 24-32): “przyjęcie do wiadomości informacji w sprawie zamiaru zawarcia umowy doradztwa w zakresie budowy zintegrowanego modelu obsługi Contact Center Gen AI z McKinsey & Company Polska sp. z o.o. sp. k.” (taking note of the plan to sign an advisory contract with McKinsey for an integrated Gen AI contact-centre model).
The closest of the seventeen to board oversight of AI: the board saw the AI Policy as part of the risk management system it oversees. It took note of the policy. The report does not say that the board approved it or that it oversees AI.
Budimex
The only rule that can end an AI tool. The Innovation Office measures the economic impact of each AI tool it tests and recommends keeping or dropping it. The report does not say who makes the final call. The management boards answer for a compliance system whose risk areas include the use of AI. The supervisory board's report does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: Budimex S.A.
Read: Sprawozdanie Zarządu z działalności Grupy Budimex i Budimex SA w 2025 roku, PDF, 192 pages, dated 26 March 2026, in the package budimex.pl, listed at budimex.pl. The consolidated financial statements and the CEO's letter in the same package do not mention AI.
Who decides about AI?YesPage 183.
“Biuro Innowacji prowadzi projekty oparte o wykorzystanie różnorodnych modeli językowych i systemów wizyjnych. Rozwiązania wykorzystujące technologie AI podlegają tym samym zasadom oceny co inne inicjatywy wdrożeniowe – w szczególności Biuro Innowacji prowadzi rygorystyczne pomiary ich wpływu ekonomicznego oraz rekomenduje dalsze stosowanie bądź rezygnację z testowanych narzędzi.”
The Innovation Office runs projects that use a range of language models and vision systems. Solutions that use AI technologies are subject to the same assessment rules as other implementation initiatives. In particular, the Innovation Office rigorously measures their economic impact and recommends either continuing to use the tested tools or dropping them.
The Innovation Office recommends. The report does not say who takes the final decision.
Does the board oversee AI?PartialPages 179-180.
“Główne obszary szczegółowych ryzyk prawnych, których wystąpieniu zapobiega system compliance w Grupie Budimex, to: ... wykorzystywanie sztucznej inteligencji.”
The main areas of specific legal risk that the compliance system in the Budimex Group prevents are: ... the use of artificial intelligence.
“Zgodnie z obowiązującą w Grupie Budimex Polityką compliance organami odpowiedzialnymi za wprowadzenie w Spółkach Polityki compliance, aktualizowanie systemu compliance, wdrażanie procedur przestrzegania zasad wyrażonych w Polityce compliance oraz zawartych w systemie compliance, jak również wynikających z obowiązujących przepisów prawa są Zarządy Spółek.”
Under the Compliance Policy in force in the Budimex Group, the bodies responsible for introducing the Compliance Policy in the companies, updating the compliance system and implementing procedures to observe the rules of the Compliance Policy, of the compliance system and of applicable law are the companies' management boards.
The management boards answer for a compliance system whose risk areas include the use of AI. Both passages sit in one section, on the compliance system. The report does not state supervisory-board oversight of AI.
Is there a rule that can stop AI?YesPage 183.
“Rozwiązania wykorzystujące technologie AI podlegają tym samym zasadom oceny co inne inicjatywy wdrożeniowe – w szczególności Biuro Innowacji prowadzi rygorystyczne pomiary ich wpływu ekonomicznego oraz rekomenduje dalsze stosowanie bądź rezygnację z testowanych narzędzi.”
Solutions that use AI technologies are subject to the same assessment rules as other implementation initiatives. In particular, the Innovation Office rigorously measures their economic impact and recommends either continuing to use the tested tools or dropping them.
“Grupa rozwija rozwiązania w oparciu o pilotaże oraz systematyczną ewaluację wdrażanych technologii.”
The Group develops solutions through pilots and a systematic evaluation of the technologies it deploys.
Economic impact is the only criterion the report names.
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
CD PROJEKT
A management board member owns AI. The Joint Chief Operating Officer oversees technological innovation, including AI tools and skills, and AI research reports directly to that member. The AI team narrowed its research projects, but the report states no rule for stopping an AI tool. The supervisory board's report does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: CD PROJEKT S.A.
Read: Sprawozdanie Zarządu z działalności Grupy Kapitałowej CD PROJEKT za okres od 1 stycznia do 31 grudnia 2025 roku, PDF, 156 pages, March 2026, cdprojekt.com. The consolidated financial statements do not mention AI.
Who decides about AI?YesPage 68, governance statement, on management board member Paweł Zawodny.
“Joint Chief Operating Officer, Członek Zarządu ... Nadzoruje działania firmy w obszarze innowacji technologicznych, w tym rozwój narzędzi i kompetencji związanych ze sztuczną inteligencją (AI). Bezpośrednio podlegają mu obszary: Technologia | Badania i rozwój (R&D), w tym sztuczna inteligencja (AI)”
Joint Chief Operating Officer, Member of the Management Board ... Oversees the company's work in technological innovation, including the development of tools and skills related to artificial intelligence (AI). Areas reporting directly to him: Technology, Research and development (R&D), including artificial intelligence (AI).
Does the board oversee AI?YesPages 68 and 74.
Page 68, the same management board member:
“Nadzoruje działania firmy w obszarze innowacji technologicznych, w tym rozwój narzędzi i kompetencji związanych ze sztuczną inteligencją (AI).”
Oversees the company's work in technological innovation, including the development of tools and skills related to artificial intelligence (AI).
Page 74, repeated on page 88:
“Członkowie Zarządu omawiają kwestie związane ze zrównoważonym rozwojem podczas cyklicznych spotkań z przedstawicielami działów, w których kompetencjach znajdują się dane zagadnienia. W 2025 r. spotkania te dotyczyły m.in. wyników badania zaangażowania osób zatrudnionych, wdrożenia postanowień dyrektywy o jawności wynagrodzeń, nawiązywania relacji biznesowych, nowych technologii, w tym sztucznej inteligencji oraz bezpieczeństwa IT.”
Members of the Management Board discuss sustainability matters at regular meetings with representatives of the departments responsible for them. In 2025 these meetings covered, among other things, the results of the employee engagement survey, implementation of the pay transparency directive, establishing business relationships, new technologies including artificial intelligence, and IT security.
The report does not state oversight of AI by the supervisory board or its audit committee.
Is there a rule that can stop AI?PartialPage 26.
“W 2025 roku zespół odpowiedzialny za obszar Artificial Intelligence (sztucznej inteligencji) kontynuował badanie możliwości wykorzystania istniejących oraz budowy własnych narzędzi opartych o AI w rozwoju przyszłych produktów ... Część prowadzonych inicjatyw została poddana testom oraz wstępnemu wykorzystaniu. Jednocześnie zespół dokonał selekcji i konsolidacji projektów badawczych, w wyniku których podjął nowe inicjatywy, koncentrując się na najbardziej perspektywicznych kierunkach.”
In 2025 the team responsible for Artificial Intelligence continued to explore using existing AI-based tools and building its own for the development of future products ... Some of the initiatives were tested and put into initial use. At the same time, the team selected and consolidated its research projects and, as a result, took up new initiatives, concentrating on the most promising directions.
The AI team reviewed its research portfolio. The report states no rule for stopping an AI tool.
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
Dino Polska
No mention of AI. The annual report's hits for “sztuczn-” are plastics and artificial sausage casings. The supervisory board's report, in two documents, does not mention AI either.
Read Annual report 2025 · Supervisory board’s report on 2025, part 1 · Supervisory board’s report on 2025, part 2
Show the evidence, with pages
Company: Dino Polska S.A.
Read: Sprawozdanie z działalności Dino Polska i Grupy Dino Polska w 2025 r., published in XHTML only, 106 pages, approved 26 March 2026, in the package grupadino.pl, listed at grupadino.pl. The consolidated financial statements do not mention AI.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
The report does not mention AI. The nine hits for sztuczn- refer to plastics or artificial sausage casings (pages 56, 67, 78-80, 94). The risk factor on IT systems (page 28) does not mention AI.
The supervisory board’s own report on 2025
Neither of its two documents mentions AI. It was searched in full for the same AI terms as the annual report.
Erste Bank Polskaex-Santander
One AI test, no owner. The report, published as Santander Bank Polska, lists a model risk committee with no link to AI and names the fast growth of AI as context. One HR chatbot was tested and approved for rollout, with no general rule for stopping an AI tool. The supervisory board's report does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: Erste Bank Polska S.A., formerly Santander Bank Polska S.A. The FY2025 report was published under the former name.
Read: Sprawozdanie Zarządu z działalności Grupy Kapitałowej Santander Bank Polska S.A. w 2025 roku, PDF, 338 pages, published 24 February 2026, erste.pl, listed at erste.pl. The consolidated financial statements do not mention AI.
Who decides about AI?Not foundThe list of management committees (page 178) includes a model risk management committee, with no link to AI.
“Komitet Zarządzania Ryzykiem Modeli”
Model Risk Management Committee
Does the board oversee AI?Not foundThe governance statement (page 179) lists the fast development of AI among the conditions of 2025. That is context, not oversight.
“zmiany w składzie organów Banku i inne zmiany kadrowe oraz bardzo szybki rozwój systemów i modeli sztucznej inteligencji.”
changes in the composition of the Bank's governing bodies and other staff changes, and the very fast development of artificial intelligence systems and models.
Is there a rule that can stop AI?PartialPage 48.
“Opracowano asystenta chatbotowego, tj. prototypowe rozwiązanie AI mające na celu wsparcie obsługi ogólnych i powtarzalnych zapytań HR ze strony pracowników. Zakończono pierwsze testy z udziałem ekspertów HR, w oparciu, o które podjęto decyzję o wdrożeniu asystenta w I kwartale 2026 r.”
A chatbot assistant was developed, a prototype AI solution to support the handling of general and repetitive HR queries from employees. The first tests with HR experts were completed, and on that basis a decision was taken to deploy the assistant in the first quarter of 2026.
One tool went through a test and a decision to deploy. The report states no general rule and no stop criterion.
The supervisory board’s own report on 2025
It was published before the change of name, as Santander Bank Polska’s, and it does not mention AI. It was searched in full for the same AI terms as the annual report.
Grupa Kęty
No mention of AI in the annual report, its official version or the separate sustainability report. The supervisory board's report does not mention AI either.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: Grupa Kęty S.A.
Read: Sprawozdanie Zarządu Grupy Kęty S.A. z działalności Spółki i Grupy Kapitałowej w 2025 Roku, PDF marked as unofficial, 189 pages, grupakety.com, listed at grupakety.com. Also searched: the official XHTML version, and the separate sustainability report of May 2026 (88 pages), grupakety.com.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
No document read mentions AI. All hits for sztuczn- refer to plastics. Automatyzacja and robotyzacja refer to production lines.
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
KGHM
AI as a tool, not a governance topic. The report describes AI in optimising production, in a training project and in a planned research programme in human resources. The risk section and the governance statement do not mention it. The supervisory board's report does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: KGHM Polska Miedź S.A.
Read: Sprawozdanie Zarządu z działalności KGHM Polska Miedź S.A. oraz Grupy Kapitałowej KGHM Polska Miedź S.A. w 2025 r., PDF, 316 pages, kghm.com. The consolidated financial statements do not mention AI.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
AI appears as a technology in use: optimising production (page 31), a training project and a planned research programme in human resources (pages 272-273). The risk section and the governance statement do not mention AI.
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
KRUK
A management board member owns AI. The Chief Data & Technology Officer oversees AI & Digital Enablement and the development of AI services. An AI Governance project prepared the company for the AI Act with risk-assessment criteria, but the report does not say what happens to a tool that fails them. The supervisory board's report lists AI risk work among the year's tasks; it does not say that the board itself looked at AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: KRUK S.A.
Read: Sprawozdanie Zarządu z działalności Grupy KRUK i KRUK S.A. za 2025 rok, PDF, 310 pages, pl.kruk.eu, listed at pl.kruk.eu. Printed page numbers are 7 lower than PDF page numbers up to about PDF page 134, and 8 lower from about PDF page 181. The consolidated financial statements do not mention AI.
Who decides about AI?YesPages 108 and 112 (PDF pages 115 and 119), governance statement.
“Członek Zarządu pełniący funkcję Chief Data & Technology Officer (CDTO) sprawuje nadzór nad IT, Debt Portfolio Valuation, Statistical Methods Development, Cybersecurity, IT International Procurement, Core System i AI & Digital Enablement.”
The Management Board Member serving as Chief Data & Technology Officer (CDTO) oversees IT, Debt Portfolio Valuation, Statistical Methods Development, Cybersecurity, IT International Procurement, Core System and AI & Digital Enablement.
“Członek Zarządu- Chief Data & Technology Officer ... sprawuje nadzór nad rozwojem narzędzi cyfrowych i usług AI w GK KRUK”
Management Board Member, Chief Data & Technology Officer ... oversees the development of digital tools and AI services in the KRUK Group
Does the board oversee AI?YesPage 112 (PDF page 119), the same management board member's duties.
“sprawuje nadzór nad rozwojem narzędzi cyfrowych i usług AI w GK KRUK”
oversees the development of digital tools and AI services in the KRUK Group
The report does not state oversight of AI by the supervisory board or a board committee.
Is there a rule that can stop AI?PartialPage 173 (PDF page 181) and page 88 (PDF page 95).
“Zrealizowano również projekt AI Governance, przygotowując organizację do wymogów AI Act. Opracowano ujednolicone podejście do oceny ryzyka, kryteria, scenariusze oraz dokumentację zapewniającą zgodne z RODO stosowanie narzędzi AI.”
An AI Governance project was also completed, preparing the organisation for the requirements of the AI Act. A unified approach to risk assessment was developed, with criteria, scenarios and documentation to ensure that AI tools are used in line with the GDPR.
“Zdefiniowane ramy zarządzania ryzykiem w zakresie stosowania usług sztucznej inteligencji są obecnie przenoszone przez KRUKa na poziom procedur operacyjnych.”
KRUK is now carrying its defined framework for managing the risk of using artificial intelligence services down to the level of operating procedures.
The report describes AI risk assessment and criteria. It does not say what happens to a tool that fails the assessment.
The supervisory board’s own report on 2025
Sprawozdanie z działalności Rady Nadzorczej for 2025, PDF, 48 pages, attached to item 6 of the draft resolutions for the general meeting of 26 May 2026, bonnier.pl. Page 16, in the board's account of risk management, internal control, compliance and internal audit:
“W 2025 r. kontynuowane były działania na rzecz rozwoju systemu zarządzania ryzykiem i systemu kontroli wewnętrznej, funkcji compliance oraz funkcji audytu wewnętrznego” [...] “prowadzono prace nad wdrożeniem wymagań dla procesów zarządzania ryzykiem w zakresie stosowania usług sztucznej inteligencji, które będą kontynuowane w 2026 roku”
In 2025 work continued on developing the risk management system and the internal control system, the compliance function and the internal audit function [...] work was carried out on implementing requirements for risk-management processes covering the use of artificial intelligence services, to be continued in 2026.
The board describes work the company did. It does not say that the board itself looked at AI. The report also repeats that a management board member, the Chief Data & Technology Officer, “sprawuje nadzór nad IT, Debt Portfolio Valuation, Statistical Methods Development, Cybersecurity, IT International Procurement, Core System i AI & Digital Enablement” (oversees IT [...] and AI & Digital Enablement), which matches the annual report. AI also appears in sponsorship (page 20) and in a board member's biography (page 32).
LPP
AI as business use only: design, logistics, customer service and a recommendation engine. The governance and sustainability statements do not mention AI, and neither does the supervisory board's report.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: LPP S.A.
Read: GK LPP SA Skonsolidowany roczny raport za 2025, financial year 1 February 2025 to 31 January 2026, dated 26 March 2026, PDF marked as not the official version, 247 pages, lppsa.com.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
AI appears as business use on pages 5, 9, 27, 32 and 43: in the CEO's letter, in design, logistics and customer service, and in a recommendation engine. One line in the risk section (page 32) names AI as a support in design. The governance and sustainability statements do not mention AI.
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
mBank
The closest annual report. The supervisory board chair's letter puts the board's continuous oversight of operational resilience next to the safe use of AI, without saying that the board oversees AI. The report describes testing, validating and scaling generative AI, with no criterion for stopping it, and names no owner of AI. The supervisory board's own report mentions AI only in a diversity policy.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: mBank S.A.
Read: the annual report package mbank.pl, listed at mbank.pl, all eleven PDFs searched. Quoted below: Sprawozdanie Zarządu z działalności Grupy mBanku S.A. w 2025 roku (364 pages), the supervisory board chair's letter to shareholders and the CEO's letter to shareholders. The financial statements and the supervisory board's assessments do not mention AI.
Who decides about AI?Not foundThe nearest passages: a model risk committee appears in the list of management committees (page 156) with no AI remit stated, and a table of board members' training and industry activities lists work in the Polish Bank Association's AI working group (page 199). That is industry work, not a remit inside the bank.
“Komitet ryzyka modeli”
Model risk committee
“Współpraca w ramach grupy roboczej ds. AI przy Związku Banków Polskich, w ramach której była prowadzona dyskusja m.in. na temat EU AI Act”
Cooperation in the AI working group of the Polish Bank Association, which discussed, among other things, the EU AI Act
Does the board oversee AI?PartialSupervisory board chair's letter, page 3.
“Wobec rosnących zagrożeń cybernetycznych Rada Nadzorcza utrzymuje stały nadzór nad odpornością operacyjną Banku, obejmującą rozwój systemów ochrony infrastruktury krytycznej, testowanie scenariuszy kryzysowych, współpracę z instytucjami nadzorczymi oraz wzmacnianie kultury bezpieczeństwa w organizacji. Zdolność do bezpiecznego i odpowiedzialnego wykorzystania technologii, w tym sztucznej inteligencji, pozostaje jednym z kluczowych warunków utrzymania długoterminowej wartości i zaufania interesariuszy.”
In view of growing cyber threats, the Supervisory Board maintains continuous oversight of the Bank's operational resilience, covering the development of systems that protect critical infrastructure, the testing of crisis scenarios, cooperation with supervisory institutions and the strengthening of security culture in the organisation. The ability to use technology, including artificial intelligence, safely and responsibly remains one of the key conditions for maintaining long-term value and the trust of stakeholders.
A judgement call: the letter ties the supervisory board's oversight of operational resilience to the safe use of AI in adjacent sentences of one paragraph.
Is there a rule that can stop AI?PartialPage 27, in the description of the GenAI programme, and the CEO's letter, page 3.
“Ta strategia to jeden z głównych czynników przyspieszenia rozwoju, a zorganizowany proces dla testowania, walidacji i skalowania rozwiązań jest gotowy do adaptacji w innych jednostkach i sektorach.”
This strategy is one of the main factors speeding up development, and an organised process for testing, validating and scaling solutions is ready to be adapted in other units and sectors.
“udany program inkubacji rozwiązań generatywnej sztucznej inteligencji, który już zaowocował praktycznymi zastosowaniami testowanych pomysłów”
a successful incubation programme for generative artificial intelligence solutions, which has already produced practical uses of the ideas tested
The report describes testing, validation and scaling. It states no criterion for stopping an AI tool.
The supervisory board’s own report on 2025
Sprawozdanie Rady Nadzorczej mBanku S.A. z działalności w 2025 r., PDF, 53 pages, among the documents for the general meeting of 31 March 2026, bonnier.pl. Page 44, on the bank's policy for a diverse and inclusive workplace:
“W relacjach z klientami odnosi się do dostępności usług, neutralności językowej, etycznego wykorzystania AI i przeciwdziałania stereotypom.”
In relations with customers it covers access to services, neutral language, the ethical use of AI and countering stereotypes.
The chair's letter in mBank's annual report says more about AI than the board's own report does (see above).
MODIVOex-CCC
Rules for using AI, no owner. The cybersecurity rules cover the safe use of AI tools, and the supervisory board vice-chair's biography mentions AI in retail. The report names no owner of AI, no board oversight and no gate. The supervisory board's report does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: MODIVO S.A., formerly CCC S.A., renamed in February 2026.
Read: Skonsolidowane Sprawozdanie Zarządu z działalności Grupy Kapitałowej MODIVO S.A., financial year 1 February 2025 to 31 January 2026, graphic PDF, 244 pages, modivoplatform.com, checked against the official XHTML.
Who decides about AI?Not found
Does the board oversee AI?Not foundThe nearest passage is the biography of the supervisory board's vice-chair (page 74). A biography describes experience, not oversight.
“Jego doświadczenie obejmuje także rozwój strategii klientocentrycznych oraz wykorzystanie sztucznej inteligencji w handlu.”
His experience also covers developing customer-centric strategies and using artificial intelligence in retail.
Is there a rule that can stop AI?Not foundThe nearest passage sets rules for using AI (page 237). It describes no review or gate.
“Regulacje w zakresie cyberbezpieczeństwa określa zasady bezpiecznego i odpowiedzialnego korzystania z systemów oraz narzędzi AI, zgodnie z przepisami prawa, standardami i wartościami Grupy Kapitałowej MODIVO.”
The cybersecurity regulations set out the rules for the safe and responsible use of AI systems and tools, in line with the law and with the standards and values of the MODIVO Group.
AI also appears as a climate-related opportunity, in stock optimisation (page 177).
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
ORLEN
A board committee took note of AI. The supervisory board's Security Committee took note of information on AI solutions in the group. The annual report describes an AI rollout and its risks, an AI course with an ethics module and a customer-service chatbot, but names no owner of AI, no board oversight and no rule to stop.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: ORLEN S.A.
Read: Sprawozdanie Zarządu z działalności Grupy ORLEN i ORLEN S.A. za 2025 rok, PDF, 309 pages, raportroczny.orlen.pl, linked from raportroczny.orlen.pl.
Who decides about AI?Not foundThe nearest passage: a management board member's remit includes “Transformację Cyfrową i Informatykę” (digital transformation and IT, our translation) on page 284, with no mention of AI.
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
AI appears in the sustainability statement: a rollout in 2025 with risks to personal data, business continuity and customer trust (page 242), an introductory AI course with an ethics module (page 257), and the company's comments on the draft Polish act on AI systems (page 261). A customer-service chatbot and voicebot appear on page 47.
The supervisory board’s own report on 2025
Sprawozdanie Rady Nadzorczej for 2025, PDF, 22 pages, attachment 6 to the notice of the general meeting of 9 June 2026, bonnier.pl. Pages 13-14, the board's Security Committee:
“W ramach swojej pracy w roku 2025 Komitet ds. Bezpieczeństwa:” [...] “zapoznał się z informacją w sprawie implementacji rozwiązań AI w Grupie Kapitałowej ORLEN”
In its work in 2025 the Security Committee [...] took note of information on the implementation of AI solutions in the ORLEN Group.
Pepco Group
AI only in the strategy: proofs of concept in the 2025 financial year and a rollout planned for 2026. The report discloses no owner, no oversight and no rule to stop.
Read Annual report 2025
Show the evidence, with pages
Company: Pepco Group N.V.
Read: Annual Report and Accounts 2025, financial year to 30 September 2025, dated 15 January 2026, PDF, 232 pages, pepcogroup.eu. Printed page numbers are 2 lower than PDF page numbers.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
AI appears only in the strategy section: proofs of concept in the 2025 financial year and a rollout planned for 2026 (page 26, PDF page 28).
The supervisory board’s own report on 2025
Pepco Group has a single board of directors and no supervisory board. Its board committees report inside the annual report above.
PGE
AI as a workforce risk and a source of demand, not as something anyone owns. The report discloses no owner, no oversight and no rule to stop. The supervisory board's report, a scan read with text recognition, does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025 (in a ZIP of meeting documents)
Show the evidence, with pages
Company: PGE Polska Grupa Energetyczna S.A.
Read: Sprawozdanie Zarządu z działalności PGE Polskiej Grupy Energetycznej S.A. oraz Grupy Kapitałowej PGE za rok 2025, PDF marked as not the official version, 368 pages, gkpge.pl.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
AI appears as a workforce risk, where AI and robotics may reduce jobs (page 290), in the distribution strategy (page 248), as tools and skills (page 216) and as a driver of demand from data centres (page 31).
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
PKO BP
An AI Competence Centre runs AI. It has begun introducing rules for the oversight and security of AI systems, including model risk classification. The report does not say to whom the centre reports or what would stop a model, and no board is said to oversee AI. The supervisory board's Strategy and IT Committee took note of information on the bank's AI competence.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: PKO Bank Polski S.A.
Read: Sprawozdanie Zarządu z działalności Grupy Kapitałowej PKO Banku Polskiego S.A. za 2025 rok, PDF, 280 pages, pkobp.pl, linked from the page of the general meeting of 29 June 2026.
Who decides about AI?YesPage 9.
“Rozwój AI jest prowadzony w ramach Centrum Kompetencji AI, które rozpoczęło wprowadzanie zasad nadzoru i bezpieczeństwa systemów AI obejmujące klasyfikację ryzyka modeli, zarządzanie danymi oraz przejrzystość, a platformy AI zostały objęte wymaganymi procesami bezpieczeństwa i odporności ICT.”
AI development is run within the AI Competence Centre, which has begun introducing rules for the oversight and security of AI systems, covering model risk classification, data management and transparency, and the AI platforms have been brought under the required ICT security and resilience processes.
The report does not say to whom the centre reports.
Does the board oversee AI?Not foundThe nearest passage is a management board member's biography (page 149), which lists AI among her skills. A biography describes skills, not oversight.
“W ostatnich latach intensywnie rozwijała kompetencje w zakresie zarządzania IT, nowoczesnych metod prowadzenia projektów (AGILE, SCRUM, DevOps) oraz technologii takich jak chmura obliczeniowa (AWS, Google Cloud) i sztuczna inteligencja.”
In recent years she has built up her skills in IT management, modern project methods (AGILE, SCRUM, DevOps) and technologies such as cloud computing (AWS, Google Cloud) and artificial intelligence.
Is there a rule that can stop AI?PartialPages 9-10.
“zasad nadzoru i bezpieczeństwa systemów AI obejmujące klasyfikację ryzyka modeli, zarządzanie danymi oraz przejrzystość”
rules for the oversight and security of AI systems, covering model risk classification, data management and transparency
“platformy MLOps/LLMOps oraz AI Gateway, umożliwiające bezpieczne i skalowalne wdrażanie modeli AI w procesach Banku.”
MLOps/LLMOps platforms and an AI Gateway, enabling secure and scalable deployment of AI models in the Bank's processes.
The report describes risk classification and secure deployment of AI models. It states no criterion for stopping one.
The supervisory board’s own report on 2025
Sprawozdanie Rady Nadzorczej PKO Banku Polskiego SA za 2025 rok, PDF, 77 pages, attached to the notice of the general meeting of 29 June 2026, bonnier.pl. Pages 64-65, the annual report of the “Komitet ds. Strategii i Technologii Informacyjnej Rady Nadzorczej”:
“W ramach realizacji zadań związanych ... wspieraniem Rady Nadzorczej w nadzorowaniu zarządzania ryzykiem związanym z technologią informacyjną Komitet” [...] “Jednocześnie Komitet zapoznał się z informacją dotyczącą budowy ... kompetencją Sztucznej Inteligencji (AI) i obszarem nowych technologii.”
In its tasks of supporting the Supervisory Board in overseeing the management of IT risk, the Committee [...] At the same time, the Committee took note of information on building, developing and managing the Artificial Intelligence (AI) competence and the area of new technologies. (the PDF's text layer splits two words, "ze" and "rozwoju", so the quotation skips them)
Elsewhere, in the board's account of the bank's year, AI is context: a strategic priority, “nowoczesną technologię (w tym AI) wspierającą biznes” (modern technology, AI included, supporting the business), and among the year's results, “wdrożyła szereg narzędzi AI wspierających pracowników, automatyzujących procesy oraz usprawniających obsługę klientów” (deployed a range of AI tools that support employees, automate processes and improve customer service). Page 75 names AI among sponsorship themes.
PZU
AI sits with the management board. One management board member's remit includes AI, a management committee coordinates data and AI governance, and in 2025 the management board adopted or updated a Data & AI Governance Policy. The report describes monitored testing but no rule for stopping an AI tool. The supervisory board's own report on 2025 does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: Powszechny Zakład Ubezpieczeń S.A. (PZU).
Read: Sprawozdanie Zarządu z działalności Grupy Kapitałowej PZU i PZU SA za rok zakończony 31 grudnia 2025 roku (Raport Roczny 2025), PDF, 418 pages, with the CEO's letter and the supervisory board chair's letter, pzu.pl.
Who decides about AI?YesPages 224 and 219, governance statement.
Page 224, the table of management committees:
Committee: “Komitet Data & AI Governance”. Purpose: “Koordynacja w PZU procesów związanych z zarządzaniem Data & AI Governance”
Committee: Data & AI Governance Committee. Purpose: coordination of the processes in PZU related to Data & AI Governance management.
“Komitety działające w PZU są kolegialnymi organami powoływanymi i likwidowanymi przez Zarząd. Wydają opinie oraz podejmują decyzje w sprawach określonych w ich regulaminach.”
The committees operating in PZU are collective bodies appointed and dissolved by the Management Board. They issue opinions and take decisions on the matters set out in their terms of reference.
Page 219, a management board member's responsibilities:
“Jan Zimowicz Członek Zarządu PZU ... Zakres odpowiedzialności: ... innowacje i AI”
Jan Zimowicz, Member of the PZU Management Board ... Scope of responsibility: ... innovation and AI
Does the board oversee AI?YesPages 226-227, governance statement.
“Zarząd podjął inne uchwały w sprawie przyjęcia i aktualizacji różnych polityk w PZU i PZU Życie, m.in.: ... Polityki Data & AI Governance w PZU i PZU Życie”
The Management Board adopted other resolutions on adopting and updating various policies at PZU and PZU Życie, including: ... the Data & AI Governance Policy at PZU and PZU Życie
The management board adopted or updated the Data & AI Governance Policy in 2025. The supervisory board chair's letter (page 8) mentions AI but does not state oversight of it.
Is there a rule that can stop AI?PartialPages 41-42.
“Zespół GPT Lab stosuje procedury testowania innowacyjnych rozwiązań i działa na specjalnie przeznaczonych zasobach chmurowych, pochodzących od zaufanego dostawcy tych technologii. Proces jest stale monitorowany, a odpowiednie podejście do testowania sztucznej inteligencji, stanowi gwarancję bezpiecznego i wiarygodnego przeprowadzania eksperymentów z technologią.”
The GPT Lab team applies procedures for testing innovative solutions and works on dedicated cloud resources from a trusted provider of these technologies. The process is continuously monitored, and the right approach to testing artificial intelligence guarantees that experiments with the technology are run safely and reliably.
“Spośród ukończonych 13 pilotaży, 7 zostało przekazanych do fazy wdrożeniowej.”
Of the 13 pilots completed, 7 were passed on to the deployment phase.
The pilots cover innovative solutions in general, AI among them. The report names no criterion for stopping an AI tool or ending a pilot.
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
TAURON
AI as a service tool: customer channels, resilience measures and a language tutor for staff. The report discloses no owner, no oversight and no rule to stop. The supervisory board's report, a scan read with text recognition, does not mention AI.
Read Annual report 2025 · Supervisory board’s report on 2025
Show the evidence, with pages
Company: TAURON Polska Energia S.A.
Read: Sprawozdanie Zarządu z działalności TAURON Polska Energia S.A. oraz Grupy Kapitałowej TAURON za 2025 r., the official XHTML, 456 pages, in the ESEF package tauron.pl. The CEO's letter, the supervisory board's assessment and the audit committee's statement in the same package do not mention AI.
Who decides about AI?Not found
Does the board oversee AI?Not found
Is there a rule that can stop AI?Not found
AI appears as a service tool: digital customer channels (page 27), resilience measures (page 232), an AI language tutor for staff (page 366) and new customer-service technologies (page 436).
The supervisory board’s own report on 2025
It does not mention AI. It was searched in full for the same AI terms as the annual report.
Żabka
A business unit owns AI tools. Żabka Future develops AI-driven technology tools for the group. The report lists AI risks and an AI policy with risk procedures, but no single passage says that a board oversees AI, and none names a rule to stop an AI tool.
Read Annual report 2025
Show the evidence, with pages
Company: Żabka Group S.A.
Read: Zabka Group S.A. Annual Report 2025, in English, the binding language version, PDF, 291 pages, zabkagroup.azureedge.net. The Polish translation was also read: zabkagroup.azureedge.net. Printed page numbers are 2 lower than PDF page numbers.
Who decides about AI?YesPage 46 (PDF page 48), corporate governance.
“Żabka Future, managed by Wojciech Krok, oversees the Group’s eGrocery and direct-to-consumer (D2C) meal solutions business. It is also responsible for the development of the Żabka Nano network and AI-driven technology tools that support Żabka Polska and other Group processes.”
Does the board oversee AI?Not foundThe report does not disclose it in one passage. It lists three AI-related emerging risks (pages 57-58, PDF pages 59-60). A separate passage (page 52, PDF page 54) says how emerging risks are reported to committees. Under the rule, a link pieced together from separate passages is not scored.
“Weakness of data governance driven by AI including unintended data feeding AI models, causing inaccurate results, legal or policy breaches and privacy failures, and AI shadowing.”
“Identified emerging risks are reported to senior management as they arise and quarterly to our Risk & Compliance Committee. At least once a year we report the results of our risk management and internal control processes to the Management Committee and Board Audit Committee”
Is there a rule that can stop AI?PartialPage 61 (PDF page 63), among the actions taken in 2025 to adapt to new laws.
“we put our AI policy into effect. It includes specific risk management and data privacy procedures and ensures our compliance with the AI Act”
The policy includes risk management procedures specific to AI. The report states no criterion for stopping an AI tool.
The supervisory board’s own report on 2025
Żabka has a single board of directors and no supervisory board. Its board committees report inside the annual report above.
Judgement calls
Where readings differed, or a call was close, the rule below decided:
- Bank Pekao, who decides: PARTIAL. A cross-check reading scored it YES. The report names a department after AI but does not describe its remit.
- Żabka, board oversight: NOT FOUND, not PARTIAL. The AI risks (pages 57-58) and their reporting to committees (page 52) sit in separate passages. Alior Bank's board oversight is NOT FOUND for the same reason.
- mBank, board oversight: PARTIAL, a judgement call. The supervisory board chair's letter ties its oversight of operational resilience to the safe use of AI in adjacent sentences of one paragraph.
- mBank, PKO BP, PZU and Żabka, rule to stop: PARTIAL. Each report describes a review or gate specific to AI, with no stop outcome.
How the reports were read
- Companies. The twenty companies in the WIG20 portfolio as of the session of 2 October 2026, as published by GPW Benchmark (Portfel tab).
- Annual reports. Each company’s annual report for the financial year 2025: the management report with its corporate governance statement and its sustainability statement, plus the letters and other documents named under each company. Pepco Group’s financial year ended on 30 September 2025; LPP’s and MODIVO’s ran from 1 February 2025 to 31 January 2026. Downloaded and read on 5 October 2026.
- Supervisory boards’ reports. Every supervisory board of a Polish joint-stock company gives the general meeting a written report on its year, with its assessment of the company’s internal control, risk management, compliance and internal audit (art. 382 § 3 pkt 3 and § 3¹ of the Commercial Companies Code). The seventeen reports on 2025 were read on 31 October 2026, from the materials for the 2026 general meetings. Alior Bank’s, PGE’s and TAURON’s are scans and were read with text recognition. Allegro.eu, Pepco Group and Żabka Group have a single board of directors and no supervisory board.
- Search. Each document was searched in full for AI terms in Polish and English, among them sztuczna inteligencja, AI, uczenie maszynowe, machine learning, generative AI, language models and the AI Act. Every hit was read in context. Hits for plastics and other false matches were set aside.
- Not read. The banks’ Pillar 3 reports, standalone policies and company websites. These could say more.
- Versions, pages and quotes. Where a company publishes a PDF, the PDF was read. Allegro.eu, Grupa Kęty, LPP, MODIVO and PGE mark their PDF as not the official version; for Grupa Kęty and MODIVO the official XHTML was checked as well. Page numbers are the numbers printed on the page; where the PDF page differs, it is given too. Passages are quoted verbatim in the language of the report. Polish passages carry our English translation, and every passage was checked against the saved text of the report.
The three questions and the rule
One rule applies to all twenty annual reports. A score rests on what one passage says. A link to AI pieced together from separate passages is not scored.
- Who decides about AI? YES when the report names a role, committee or unit and states that AI is part of its remit. PARTIAL when the answer is indirect: a unit named after AI with no remit described, a board approving an AI policy along with all other policies, an advisory AI role, or a broader committee that the same passage ties to machine learning.
- Does the board oversee AI? YES when the report says that the management board, one of its members, the supervisory board or a board committee oversees AI or AI risk. PARTIAL when a board oversees something that the same passage ties to AI, such as a code of conduct, a risk analysis or a compliance system. For the three companies registered abroad, which have a single board of directors, that board counts.
- Is there a rule that can stop AI? YES when the report states a rule, criterion, gate or review whose outcome can be to stop, drop or withdraw an AI tool or system. PARTIAL when the report describes a review, test or gate specific to AI but says nothing about stopping.
- NOT FOUND means the report does not disclose it.
Corrections
If a report says more than this sheet shows, write to pm@piotrmechlinski.com with the page, and the sheet will be corrected with the date of the change.