AI body and review procedure documented, but no AI oversight on the board
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
The 20-F confirms a group-wide AI body and a GenAI charter, and Sanofi documents a risk procedure with around 50 assessed systems; however, the process information dates from 2024, and since the change of CEO no Chief Digital Officer is listed on the executive committee.
What is documented
- The 2025 20-F confirms a group-wide governance body for AI initiatives and a GenAI charter for employees1
- Sanofi documents a Responsible AI Working Committee since April 2023, a body for high-risk applications and an internal AI policy in the employee and supplier codes4 13
- The AI Risk Assessment classifies applications by purpose, data, model and use; by May 2024 around 50 systems had been submitted and around 20 high-risk cases reviewed4
- The principles require risk-based controls also for third-party AI and training of all employees in responsible AI9
- The externally operated compliance helpline recorded 1,095 reports and 333 substantiated cases in 202511
The case against
Oversight at the top unclear
Sources for Responsibility (13)
- Sanofi: Annual Report on Form 20-F 2025 (fiscal year ended 2025-12-31) · 17 February 2026 · primary source
- Sanofi (SEC EDGAR, XBRL): Form 20-F FY2025, R53: Cybersecurity Risk Management and Strategy Disclosure · 17 February 2026 · primary source
- Sanofi (SEC EDGAR): Exhibit 1.2 to Form 20-F 2025: Board Charter (internal rules), approved by the board on 2025-04-30 · 17 February 2026 · primary source
- Sanofi (Magazin): All in on AI, Accountable to Outcomes · 16 May 2024 · primary source
- Sanofi: Principles on Data Ethics (September 2024, approved by the Bioethics Committee) · 1 September 2024 · secondary source
- Sanofi: Sanofi Bioethics Framework · 3 November 2026 · secondary source
- Sanofi: Executive Committee (web page, as retrieved) · 3 November 2026 · secondary source
- Sanofi: Governance (web page) · 3 November 2026 · secondary source
- Sanofi: Responsible AI Guiding Principles · 3 November 2026 · secondary source
- LeMondeInformatique: Sanofi accélère sur sa stratégie GenAI · 18 June 2025 · secondary source
- Sanofi: ESG Index: Ethics and Business Integrity · 31 December 2025 · primary source
- Sanofi: Sustainability Statement 2025 (chapter 3 of the URD 2025) · 17 February 2026 · primary source
- LeMondeInformatique: Sanofi travaille sur une IA plus éthique · 3 June 2024 · secondary source
2 · Safety
The 20-F documents a cybersecurity organisation based on NIST with penetration testing, a security operations centre and committee oversight; for AI itself there are only company statements on human review and monitoring.
What is documented
- Item 16K describes prevention, detection, response and recovery based on NIST, penetration testing, attack simulations, a security operations centre and supplier assessments6 7
- The Risk Committee and Audit Committee oversee cybersecurity; the CISO has been in office since 20146 7
- For Turing no action is taken without human review; risk models for trial participants have model cards and drift monitoring1 2
- According to Sanofi, ARTEMIS for adverse event reports and generative AI in regulatory documents run under human oversight4 5
- The 20-F names no material cyber incident; a 2025 hacker claim is unconfirmed7 10
The case against
AI testing only self-reported
Further gaps
- Certifications such as ISO/IEC 27001 and specific protection of the AI platforms are not documented
- No statement by Sanofi on the 2025 claim by the R00TK1T group was found10
Sources for Safety (10)
- Sanofi (Magazin): All in on AI, Accountable to Outcomes · 16 May 2024 · primary source
- Sanofi (Magazine): AI: Transparent and Explainable · 3 November 2026 · secondary source
- LeMondeInformatique: Sanofi accélère sur sa stratégie GenAI · 18 June 2025 · secondary source
- Campus Sanofi (Spain): IA en farmacovigilancia: Sanofi automatiza el análisis de efectos adversos en tiempo real (ARTEMIS) · 2 October 2025 · secondary source
- DGRA / Sanofi (Yves Jacob, Head Regulatory Europe): Presentation at the DGRA annual congress 2025 · 8 May 2025 · secondary source
- Sanofi (SEC EDGAR, XBRL): Form 20-F FY2025, R53: Cybersecurity Risk Management and Strategy Disclosure · 17 February 2026 · primary source
- Sanofi (SEC EDGAR, XBRL): Form 20-F FY2025, R55: Cybersecurity Risk Management and Strategy Disclosure (Tagging) · 17 February 2026 · primary source
- Sanofi: Sécurité de l'information et confidentialité des données (ESG Index) · 3 November 2026 · secondary source
- Sanofi (SEC EDGAR, XBRL): Form 20-F FY2025, R36: Legal and arbitral proceedings (Note D.22) · 17 February 2026 · primary source
- The Cyber Express: R00TK1T Announces Cyberattack on Sanofi, Raises Alarming Human Experimentation Claims (Tag-Seite) · 9 July 2025 · secondary source
3 · Transparency
Sanofi describes its AI applications publicly and commits to communicating their use and limits, but publishes code and model information only for the research model CodonBERT.
What is documented
- Sanofi publicly describes its AI applications in research, trials, manufacturing and sales3
- The principles commit to communicating when and how AI is used and where its limits lie, and to enabling users to reject results2
- Explainable models and feature contributions are used for risk models, and explainability statistics for Turing4 5
- CodonBERT is public with code, weights, training data IDs and licences6
The case against
Documentation remains internal
- Model cards and explainability methods are described only as internal practice and are not published5
- There are no public system cards and no transparency report for ARTEMIS, Muse and Concierge
- Labelling towards trial participants, healthcare professionals or patients is not documented
Sources for Transparency (6)
- Sanofi: Annual Report on Form 20-F 2025 (fiscal year ended 2025-12-31) · 17 February 2026 · primary source
- Sanofi: Responsible AI Guiding Principles · 3 November 2026 · secondary source
- Sanofi: Artificial Intelligence across Sanofi · 3 November 2026 · secondary source
- Sanofi (Magazin): All in on AI, Accountable to Outcomes · 16 May 2024 · primary source
- Sanofi (Magazine): AI: Transparent and Explainable · 3 November 2026 · secondary source
- Sanofi-Public (GitHub): CodonBERT README · 3 November 2026 · primary source
4 · Why the harm rule does not apply
The harm rule does not apply: within the window there is no judgment, authority decision or fine on harm from AI. The only recorded controversy is an unconfirmed 2025 hacker claim with no documented link to AI.
5 · Controversies (1)
6 · What remains open
- Governance, board and risk chapters of the 2025 20-F and URD, which were truncated on retrieval
- Responsibility for AI and digital after the 2026 change of leadership
- Current figures on AI risk reviews and training after 2024
- Validation and test results for AI in drug safety and trials
- Approval procedure for models purchased from external providers
- Chapters on data protection and governance in the 2025 sustainability statement
7 · How it was assessed
Cut-off date: 3 November 2026 · 41 searches, 37 fetches.
Primary documents read (10)
- Form 20-F FY2025 (Item 3.D Risk Factors pp. 1–14) and XBRL R36, R53, R55
- Exhibit 1.2 Board Charter (2025-04-30)
- Document d'enregistrement universel 2025 (up to p. 14)
- Sustainability Statement 2025 (pp. 1–27)
- Minutes of the Annual General Meetings 2025 and 2026
- Responsible AI Guiding Principles
- Principles on Data Ethics (September 2024)
- Sanofi Bioethics Framework
- ESG Index pages: Ethics and Business Integrity and Sécurité de l'information et confidentialité des données
- Press releases dated 12 February 2026 and 17 February 2026
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0