AI principles and Supervisory Board oversight, but little evidence of implementation
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
ASML documents concrete general governance and AI principles, but the Responsible AI Policy is not published, and risk classes, approval procedures and AI figures are missing.
What is documented
- The Audit Committee oversees risk management and the use of ICT; a CESR committee oversees risk management2 3
- In 2025 the Supervisory Board discussed AI opportunities in a deep dive and supports the AI strategy; training on AI is planned1
- ASML names nine AI principles and a group-wide Responsible AI Policy; a Privacy & AI Compliance Office is building an AI compliance programme5 4
- The Code of Conduct contains rules on generative AI; the Speak Up service can be used anonymously and is also open to third parties8
The case against
Principles without documented implementation
No figures or responsibility
- No committee or member of executive management is explicitly responsible for AI risks2
- AI figures, an AI report and an external audit of AI governance are missing
Sources for Responsibility (10)
- ASML: 2025 Annual Report – Corporate governance section · 25 February 2026 · primary source
- ASML: Rules of Procedure Supervisory Board (version of 22 April 2026) · 22 April 2026 · primary source
- ASML: Risk (Governance) · 3 November 2026 · primary source
- ASML Careers: Legal and compliance internship: Privacy & AI Compliance Office · 3 November 2026 · secondary source
- ASML: Responsible AI · 3 November 2026 · secondary source
- ASML Careers: Ethics, Privacy and AI Officer (Chandler, USA) · 3 November 2026 · secondary source
- ASML (investor.asml.com): ASML, Mistral AI enter strategic partnership · 9 September 2025 · primary source
- ASML: Code of Conduct (version 1.0, valid from October 2024) · 3 November 2026 · primary source
- ASML (investor.asml.com): Reporting on sustainable long-term value creation (Annual Report 2025, Sustainability) · 31 December 2025 · primary source
- ASML: Voting Results 2026 Annual General Meeting of ASML Holding NV · 22 April 2026 · primary source
2 · Safety
ASML documents IT security reporting to the Supervisory Board, but certifications, testing and incident processes are barely documented, and an espionage case points to gaps in insider protection.
What is documented
- The Supervisory Board receives regular reports on IT security; in 2025 the Audit Committee reviewed the Risk, Business Assurance & Security function4 5
- A job advertisement mentions an AI security strategy and alignment with NIST CSF, ISO 27001, IEC 62443 and NIS26
- After an investigation, ASML rejected a leak claim from January 2026 as unfounded11 12
The case against
Protection against data leakage
AI testing
- There is no information on validation, monitoring and limitations of the AI in its systems and diagnostics1
Sources for Safety (13)
- ASML: 2025 Annual Report – Strategic report section · 25 February 2026 · secondary source
- ASML: Responsible AI · 3 November 2026 · secondary source
- Mistral AI: Customer story ASML · 3 November 2026 · secondary source
- ASML: 2025 Annual Report – Corporate governance section · 25 February 2026 · primary source
- ASML: Rules of Procedure Supervisory Board (version of 22 April 2026) · 22 April 2026 · primary source
- ASML Careers: Senior Security AI Specialist · 3 November 2026 · secondary source
- ASML Careers: Security manager compliance · 3 November 2026 · secondary source
- The Register: Ex-ASML engineer who stole chip tech for Russia gets three years in Dutch prison · 10 July 2025 · secondary source
- Omroep Brabant: Rus deelt bedrijfsgeheimen van ASML en gaat drie jaar cel in · 10 July 2025 · secondary source
- NL Times: Suspected Russian spy says ASML secrets were to maintain knowledge and skills · 3 April 2025 · secondary source
- Silicon UK: ASML says hack claims are false · 8 January 2026 · secondary source
- Newsquawk: ASML (ASML NA) says 'Thorough Investigation' has shown that none of the Co.'s data was exposed · 6 January 2026 · secondary source
- ASML: Code of Conduct (version 1.0, valid from October 2024) · 3 November 2026 · primary source
3 · Transparency
ASML publishes principles on explainability and names AI applications in general terms, but information on data, limitations and models used, as well as an AI transparency report, is missing.
What is documented
- ASML publishes nine AI principles, including that AI systems should make their purpose, inputs and outputs understandable1
- The annual report names AI applications in lithography, planning and diagnostics in general terms2 3
- The privacy notice for applicants mentions the possible use of AI systems and a right to object to automated decisions4
The case against
Missing system information
No reporting
- An AI transparency report, a register of AI systems and labelling of AI-supported functions for customers are missing
Sources for Transparency (6)
- ASML: Responsible AI · 3 November 2026 · secondary source
- ASML: 2025 Annual Report – Strategic report section · 25 February 2026 · secondary source
- ASML: 2025 Annual Report based on US GAAP (full report) · 25 February 2026 · secondary source
- ASML: Privacy Notice for Job Applicants (as of October 2025) · 3 November 2026 · secondary source
- Mistral AI: Customer story ASML · 3 November 2026 · secondary source
- ASML (investor.asml.com): ASML, Mistral AI enter strategic partnership · 9 September 2025 · primary source
4 · Why the harm rule does not apply
The harm rule does not apply: there is no AI-related finding, no fine and no pattern of serious incidents with its own AI systems.
5 · Controversies (2)
- minorAccording to a report that could not be retrieved, prosecutors sought three years in prison for former ASML employees over data theft; the details are unclear. · 26 February 2026 · unklar
- minorThe Rotterdam District Court sentenced a former ASML and NXP engineer to three years in prison for copying confidential documents and passing them to Russia. · 10 July 2025 · abgeschlossen
6 · What remains open
- Text and implementation status of the Responsible AI Policy
- Responsibility for AI risks at Board of Management and Supervisory Board level
- Security certifications, testing and incident processes according to the Information security section of the 20-F
- Validation and limitations of the AI in lithography and diagnostics software
- Protection of customer and system data in the cooperation with Mistral AI
- Measures against insiders after the espionage case and status of a further case against former employees
7 · How it was assessed
Cut-off date: 3 November 2026 · 42 searches, 37 fetches.
Primary documents read (8)
- Annual Report 2025, strategic report section (S03, three retrievals) and corporate governance section (S04, two retrievals)
- Annual Report 2025, full PDF, up to p. 37 only (S07)
- Form 20-F, financial year 2025, pp. 1–46 only (S05, two retrievals)
- Annual Report 2025, sustainability section (S06)
- Responsible AI page (S01), Code of Conduct 2024 (S09), Rules of Procedure Supervisory Board 2026 (S10, two retrievals), risk-governance page (S11), Governance downloads (S08)
- Voting results, Annual General Meeting 2026 (S35); privacy notice for applicants (S36)
- Press releases: ASML/Mistral (S17) and Mistral (S16)
- Official finding: press release, House Select Committee on the CCP (S27)
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0