AI in grid software without its own AI policy and without disclosed testing
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
The board deals with AI as a strategic topic, but there is no responsibility for AI risks, no published AI policy and no approval procedure for own and acquired AI.
What is documented
- The Audit Committee reviews cybersecurity and risk management at least once a year2 1
- The board's 2025 strategy review covered AI, automation and robotics as priorities1
- The code of conduct allows only approved AI applications and prohibits confidential data in public AI tools5
- According to a blog, development follows a framework focused on low risk and human augmentation8
- General reporting channels exist, with an integrity portal, an ombudsperson and a ban on retaliation5 11
The case against
No responsibility for AI
Only principles, no procedures
- The Acceptable Use Policy and AI Guidelines are only mentioned in the sustainability report; their content is not public6
- There is no risk classification or approval procedure for own and acquired AI, such as from Alteia16
- External assurance in the sustainability report covers only greenhouse gas data; AI certifications are missing6 14
Sources for Responsibility (16)
- GE Vernova Inc. / U.S. SEC: 2026 Proxy Statement (Form DEF 14A) · 20 May 2026 · primary source
- GE Vernova Inc.: Audit Committee Charter (effective 30 September 2026) · 30 September 2026 · primary source
- Board Cybersecurity (board-cybersecurity.com): GE Vernova Inc. 10-K Item 1C. Cybersecurity – 2026-01-29 · 29 January 2026 · secondary source
- Board Cybersecurity (board-cybersecurity.com): GE Vernova – Governance Tracker · 19 March 2026 · secondary source
- GE Vernova Inc.: GE Vernova Code of Conduct · 3 November 2026 · secondary source
- GE Vernova Inc.: GE Vernova Sustainability Report 2025 · 17 June 2026 · secondary source
- GE Vernova Inc.: Sustainability Executive Summary 2025 · 3 November 2026 · secondary source
- GE Vernova Electrification Software (Jay Shah): Leading Industrial AI in the Energy Sector (blog) · 3 November 2026 · secondary source
- GE Vernova Inc.: GE Vernova whitepapers offer pragmatic approach to AI for more intelligent energy grids · 10 June 2025 · secondary source
- The Muse (Stelleninserat GE Vernova): AI Governance Project Manager – GE Vernova (job posting) · 27 May 2026 · secondary source
- GE Vernova Inc.: Human Rights Statement 2025 · 3 November 2026 · secondary source
- GE Vernova Inc.: GE Vernova’s Commitment to the Protection of Personal Information (BCR-C, web page) · 3 November 2026 · secondary source
- GE Vernova Inc.: GE Vernova BCR-C (PDF) · 3 November 2026 · secondary source
- GE Vernova Inc.: 2025 Sustainability Performance Data · 3 November 2026 · secondary source
- European Data Protection Board: Register of Binding Corporate Rules (page 23 of 25) · 3 November 2026 · secondary source
- American Public Power Association: GE Vernova to acquire firm specialising in AI computer vision, machine learning · 22 July 2025 · secondary source
2 · Safety
GE Vernova documents a certified secure development process in one business unit and functioning vulnerability handling, but there is no information on testing and monitoring of the AI functions.
What is documented
- An ISASecure SDLA certificate to IEC 62443-4-1 for Power Conversion & Storage was issued in 2026 and is valid until 20295
- The CISA advisory on Enervista UR Setup names an external reporter and a fix; GES-2026-001 includes a fix11 12
- A public form accepts vulnerability and incident reports and acknowledges them within one working day9 10
- Since 2025 GE Vernova has assigned its own CVE numbers as a CVE Numbering Authority for all business units13
- According to an account of Item 1C, the programme is based on NIST CSF and ISO 27001, with external reviews and tabletop exercises6
The case against
No AI-specific testing
Sources for Safety (18)
- GE Vernova Inc.: GE Vernova whitepapers offer pragmatic approach to AI for more intelligent energy grids · 10 June 2025 · secondary source
- GE Vernova Electrification Software (Jay Shah): Leading Industrial AI in the Energy Sector (blog) · 3 November 2026 · secondary source
- Built In (Stelleninserat GE Vernova): Sr. Director, Product Cybersecurity AI, Emerging Technology & Platforms Leader (job posting) · 2 June 2026 · secondary source
- GE Vernova Inc.: GE Vernova introduces GridOS for Transmission, new AI whitepapers · 9 June 2026 · secondary source
- ISASecure / exida: ISASecure SDLA Certificate GE Vernova Power Conversion & Storage (IEC 62443-4-1) · 2 February 2026 · primary source
- Board Cybersecurity (board-cybersecurity.com): GE Vernova Inc. 10-K Item 1C. Cybersecurity – 2026-01-29 · 29 January 2026 · secondary source
- GE Vernova Gas Power M&D Cyber Security Team: US–Iran Conflict Security Statement · 3 March 2026 · secondary source
- GE Vernova Electrification: GridBeats Cybersecurity (product page) · 5 September 2024 · secondary source
- GE Vernova Inc.: Security – Vulnerability and Incident Reporting · 3 November 2026 · primary source
- GE Vernova Gas Power: Cybersecurity – Vulnerability Response · 3 November 2026 · primary source
- CIRCL / GCVE database (reproducing CISA advisory and CNA entries): ICSA-26-048-03: GE Vernova Enervista UR Setup · 17 February 2026 · primary source
- GE Vernova Electrification: Security Advisory GES-2026-001: CSD100 · 4 March 2026 · primary source
- GE Vernova Inc.: GE Vernova receives CVE Numbering Authority designation · 21 February 2025 · secondary source
- GE Vernova Inc.: GE Vernova BCR-C (PDF) · 3 November 2026 · secondary source
- DeXpose: MetaEncryptor strikes GE Vernova Inc. · 25 September 2026 · secondary source
- Breachsense: GE Vernova Data Breach · 28 September 2026 · secondary source
- hendryadrian.com: Ransom! GE Vernova Inc. (SEP-2026) · 26 September 2026 · secondary source
- American Public Power Association: GE Vernova to acquire firm specialising in AI computer vision, machine learning · 22 July 2025 · secondary source
3 · Transparency
GE Vernova names AI products and use cases in blogs and announcements, but discloses neither limitations nor error rates, data basis or any labelling of AI-supported outputs.
What is documented
The case against
Missing information on limitations
Missing model and system information
- No model or system cards or training data information were found, although GE Vernova develops its own ML and vision models
- An inventory of the AI systems used and the origin of the foundation models in internal use are not disclosed
Sources for Transparency (8)
- GE Vernova Electrification Software (Jay Shah): Leading Industrial AI in the Energy Sector (blog) · 3 November 2026 · secondary source
- GE Vernova Inc.: GE Vernova introduces GridOS for Transmission, new AI whitepapers · 9 June 2026 · secondary source
- GE Vernova Inc. / U.S. SEC: Form 10-K for the fiscal year ended December 31, 2025 · 29 January 2026 · secondary source
- GE Vernova Inc.: 2025 Form 10-K (PDF version on gevernova.com) · 3 November 2026 · secondary source
- GE Vernova Inc.: GE Vernova releases 2025 Sustainability Report (press release) · 17 June 2026 · secondary source
- GE Vernova Inc.: GE Vernova whitepapers offer pragmatic approach to AI for more intelligent energy grids · 10 June 2025 · secondary source
- GE Vernova Software: Four Key Takeaways from the Orchestrate 2026 GridOS Event (Blog) · 18 June 2026 · secondary source
- American Public Power Association: GE Vernova to acquire firm specialising in AI computer vision, machine learning · 22 July 2025 · secondary source
4 · Why the harm rule does not apply
The harm rule does not apply: within the 36-month window there is no judgment, no authority decision, no fine and no pattern with an AI link. The only recorded controversy, an unconfirmed ransomware claim, has no documented link to AI.
5 · Controversies (1)
- minorThe ransomware group MetaEncryptor claimed to have compromised GE Vernova's network; no evidence or confirmation was found. · 25 September 2026 · unklar
6 · What remains open
- Responsibility of the board or a committee for AI risks
- Content of the Acceptable Use Policy and the AI Guidelines
- Risk classification and approval procedures for own and purchased AI
- Testing, validation and monitoring of the AI functions in grid and power plant software
- Documented limitations, explainability and data basis of the AI products
- Confirmation or denial of the claimed ransomware attack of September 2026
7 · How it was assessed
Cut-off date: 3 November 2026 · 44 searches, 44 fetches.
Primary documents read (11)
- GE Vernova Form 10-K FY2025 (SEC HTML and PDF, each up to Item 1A)
- GE Vernova 2026 Proxy Statement (DEF 14A, with read-more)
- GE Vernova Audit Committee Charter dated 30 September 2026
- Form 8-K voting results 2025 and 2026
- Form 10-Q Q2 2026 (up to approx. 100,000 characters)
- GE Vernova Code of Conduct
- GE Vernova Sustainability Report 2025 (up to approx. p. 38), Executive Summary 2025, Performance Data 2025, Human Rights Statement 2025
- GE Vernova BCR-C (website and PDF)
- GE Vernova Security/Vulnerability Response pages, Advisory GES-2026-001, Security Notice 2026-03-03
- ISASecure SDLA Certificate (exida) for Power Conversion & Storage
- CISA Advisory ICSA-26-048-03 (via GCVE database)
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0