AI oversight at board level, but little information on testing of AI medical devices
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
Johnson & Johnson documents board oversight of AI and broad AI training, but publishes only general AI principles; risk classes, approval procedures and external audits are not disclosed.
What is documented
- The board and its committees oversee the use of AI with regular updates; the Science & Technology Committee lists AI as a risk area1
- In 2025 there were almost 120,000 completions of AI learning content; all employees receive basic AI training3
- A compliance guide on generative AI was introduced in 20254
- A GenAI course is a prerequisite for using generative AI tools8
- The AI principles require human review, monitoring and bias testing6
The case against
Rules not disclosed
- No published AI policy with risk classes and no approval procedure for purchased AI systems were found
- No named person with overall responsibility for AI governance is identified
- There is no external audit or certification of AI management and no AI-specific complaints channel
Sources for Responsibility (9)
- Johnson & Johnson (SEC EDGAR): Proxy Statement 2026 (DEF 14A) · 9 March 2026 · primary source
- Johnson & Johnson (SEC EDGAR): Proxy Statement 2025 (DEF 14A) · 12 March 2025 · primary source
- Johnson & Johnson: 2025 Health for Humanity Report (consolidated) · 1 June 2026 · primary source
- Johnson & Johnson: 2025 Health for Humanity Report – Accountability & Governance · 1 June 2026 · primary source
- CDO Magazine: How J&J Global Services Scales Enterprise AI Through Intelligent Operations and Governance · 9 February 2026 · secondary source
- Johnson & Johnson (jnj.com): Doing the Right Thing: Artificial Intelligence and Ethics · 3 November 2026 · secondary source
- The Wall Street Journal (reproduced via Dow Jones on itiger.com): Johnson & Johnson Pivots Its AI Strategy · 18 April 2025 · secondary source
- MIT Chief Data Officer and Information Quality (MIT CDOIQ): Q&A with Jim Swanson, Executive Vice President & Chief Information Officer, Johnson & Johnson · 5 June 2024 · secondary source
- Johnson & Johnson (jnj.com): Our Position on Data Privacy · 1 May 2023 · secondary source
2 · Safety
Johnson & Johnson documents procedures for data and system protection and channels for reporting vulnerabilities, but validation results for its AI medical devices are missing and the TruDi case is open.
What is documented
- Item 1C describes an incident response plan, a third-party risk programme, external maturity assessments and quarterly CISO reports7
- Vulnerability disclosure procedures for IT and medical devices exist, with acknowledgement within 3 and 10 working days13 14
- The information security training rate was 97%3
- The AI-supported navigation software Monarch Quest received FDA 510(k) clearance in 20251
The case against
Missing evidence of testing
- Validation results, error rates and post-market data for the AI medical devices are not published1
- A named framework, results of the external maturity assessments and information on protecting AI models are missing
Sources for Safety (16)
- MedPath: FDA Clears J&J's AI-Enhanced Monarch Robot for Improved Lung Cancer Diagnosis · 14 March 2025 · secondary source
- Johnson & Johnson (jnj.com): Doing the Right Thing: Artificial Intelligence and Ethics · 3 November 2026 · secondary source
- Johnson & Johnson: 2025 Health for Humanity Report (consolidated) · 1 June 2026 · primary source
- Aaj English TV (reproduced via Reuters): AI-enhanced medical device faces scrutiny after injury reports · 9 February 2026 · secondary source
- AI Incident Database: Incident 1436: AI-Enabled TruDi Navigation System Was Alleged to Have Contributed to Patient's Stroke During June 2022 Sinus Procedure · 28 March 2026 · secondary source
- AI Incident Database: Incident 1528: TruDi-associated stroke during sinuplasty in May 2023 · 9 June 2026 · secondary source
- Board Cybersecurity (reproduced from SEC mandatory disclosure): Johnson & Johnson 10-K Item 1C. Cybersecurity – 2026-02-11 · 11 February 2026 · primary source
- Johnson & Johnson (SEC EDGAR): Form 10-K for the fiscal year ended December 28, 2025 · 11 February 2026 · primary source
- Johnson & Johnson: 2025 Health for Humanity Report – Accountability & Governance · 1 June 2026 · primary source
- Johnson & Johnson (jnj.com): Our Position on Information Security · 1 March 2023 · secondary source
- Infosecurity Magazine: IBM Patient Data Breach Hits Johnson & Johnson · 2 August 2023 · secondary source
- Kirkland & Ellis (reproduced via Am Law Litigation Daily): Litigator of the Week Runners-Up and Shout Outs · 25 September 2026 · secondary source
- Johnson & Johnson (jnj.com): Coordinated Vulnerability Disclosure Statement (Version 2.0) · 6 July 2026 · primary source
- Johnson & Johnson MedTech: Product Security (productsecurity.jnj.com) · 3 November 2026 · primary source
- Settlement Administrator (court-ordered notice): Skin360 Settlement Notice · 3 November 2026 · primary source
- ID Tech Wire: Kenvue Agrees to $4.7M BIPA Settlement Over Neutrogena Skin-Scanning App · 24 February 2026 · secondary source
3 · Transparency
Johnson & Johnson names transparency and explainability as a principle, but publishes neither model cards nor limitations or performance data for its AI applications.
What is documented
The case against
Missing disclosure
- No model or system cards, AI transparency report or training data information were found
- Limitations and error rates of AI functions in medical devices such as Monarch are not documented5
- There are no rules for labelling AI-generated content for patients or professionals
Sources for Transparency (6)
- Johnson & Johnson (SEC EDGAR): Form 10-K for the fiscal year ended December 28, 2025 · 11 February 2026 · primary source
- Johnson & Johnson: 2025 Health for Humanity Report (consolidated) · 1 June 2026 · primary source
- Johnson & Johnson (jnj.com): Artificial Intelligence: Advancing Healthcare from New Jersey to the World · 4 September 2026 · secondary source
- Johnson & Johnson (jnj.com): Doing the Right Thing: Artificial Intelligence and Ethics · 3 November 2026 · secondary source
- MedPath: FDA Clears J&J's AI-Enhanced Monarch Robot for Improved Lung Cancer Diagnosis · 14 March 2025 · secondary source
- Johnson & Johnson MedTech (press release): Johnson & Johnson launches the Polyphonic AI Fund for Surgery to advance data-driven healthcare · 25 June 2025 · secondary source
4 · Why the harm rule does not apply
The harm rule does not apply: the TruDi case is pending without a judgment or authority decision, the Skin360 settlement concerns Kenvue, and the ChemImage judgment concerns a contract dispute.
5 · Controversies (4)
- mediumUSD 4.7 million settlement over the Neutrogena Skin360 AI skin analysis for capturing facial data without consent; the defendant is today's Kenvue. · 24 February 2026 · Vergleich
- mediumAccording to Reuters, FDA reports rose sharply after the then subsidiary Acclarent added AI to the TruDi navigation system; two lawsuits in Texas are pending. · 9 February 2026 · open
- minorThe class action against IBM and J&J over the data breach in the Janssen CarePath database was dismissed for lack of standing. · 25 September 2026 · abgewiesen
- minorAfter a judgment that Ethicon wrongfully terminated an AI development contract, Ethicon and ChemImage reached a settlement. · 19 December 2025 · Vergleich
6 · What remains open
- Published AI policy with risk classification and approval procedures
- External audit or certification of AI management
- Validation results, limitations and post-market data for the AI medical devices
- Statement on the TruDi case and any liability or indemnity rules from the Acclarent sale
- Items 1C and 3 of the 10-K in the original, which could not be read on retrieval
7 · How it was assessed
Cut-off date: 3 November 2026 · 42 searches, 43 fetches.
Primary documents read (13)
- Form 10-K FY2025 (Item 1, Item 1A, beginning of Item 1C)
- Form 10-K FY2025 Item 1C (reproduced via board-cybersecurity.com)
- Proxy Statements 2026, 2025 and 2024
- Form 8-K Annual General Meeting 2026
- 2025 Health for Humanity Report (consolidated, including continuation) and chapter Accountability & Governance
- Doing the Right Thing: Artificial Intelligence and Ethics
- Our Position on Data Privacy
- Our Position on Information Security
- Coordinated Vulnerability Disclosure Statement
- Product Security (productsecurity.jnj.com)
- Press release: Polyphonic AI Fund
- FDA Recall Z-0051-2025
- Skin360 Settlement Notice
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0