AI oversight at board level and cyber programme documented, AI approvals and testing described only by the company
Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.
1 · Responsibility
Wells Fargo documents AI oversight in the Risk Committee and a Head of AI, but approval and review procedures for AI are described only in its own accounts and interviews.
What is documented
- According to the 2026 proxy, the Risk Committee oversees technology risks including AI as well as data and model risk1
- The full board conducts periodic AI reviews; one director holds a certificate in AI ethics and oversight1
- Since November 2025 there has been a Head of AI accountable to the CEO2 3
- The annual report names a governance framework, ongoing model validation and GenAI training for more than 90,000 employees4
- According to its own account, separate teams and independent testing groups review models before customer release, including for adverse effects5 6
The case against
Missing rules and figures
- No published AI policy, AI metrics or AI-related complaints channels were found
- Approval procedures for purchased models from external providers are not described
- The Risk Management chapter of the 2025 annual report could not be read14
Sources for Responsibility (15)
- Wells Fargo & Company (SEC EDGAR): Proxy Statement 2026 (Form DEF 14A, wfc-20260318.htm) · 18 March 2026 · primary source
- Wells Fargo & Company (SEC EDGAR): Form 10-K for fiscal year 2025 (wfc-20251231_d2.htm) · 24 February 2026 · primary source
- QuantisNow (reproducing the Wells Fargo press release): Wells Fargo Names Saul Van Beurden to Lead Artificial Intelligence · 20 November 2025 · secondary source
- Wells Fargo & Company (SEC EDGAR, Form ARS): 2025 Annual Report (wfc2025annualreport-arsfil.pdf) · 3 November 2026 · primary source
- Wells Fargo Stories: How Wells Fargo builds responsible artificial intelligence · 22 September 2026 · secondary source
- American Banker: How Wells Fargo is deploying the White House's AI Bill of Rights · 17 February 2023 · secondary source
- arXiv (via alphaXiv): Model Risk Management for Generative AI In Financial Institutions (arXiv:2503.15668) · 19 March 2025 · secondary source
- Let's Data Science (via Business Insider): Wells Fargo Trains Employees For AI Transition · 16 March 2026 · secondary source
- Wells Fargo & Company: April 2026 Investor Presentation (proxy supplement) · 3 November 2026 · secondary source
- Wells Fargo & Company (SEC EDGAR): Form 8-K, voting results of the 2026 annual meeting (wfc-20260428.htm) · 28 April 2026 · primary source
- Wells Fargo & Company (SEC EDGAR): Form 8-K, voting results of the 2025 annual meeting (wfc-20250429.htm) · 29 April 2025 · primary source
- Wells Fargo & Company (SEC EDGAR): Form 8-K, voting results of the 2024 annual meeting (wfc-20240430.htm) · 30 April 2024 · primary source
- United States District Court, Northern District of California (via Orrick InfoBytes): In re Wells Fargo Mortgage Discrimination Litigation, No. 22-cv-00990-JD, Order re Class Certification (Dkt. 328) · 5 August 2025 · primary source
- Wells Fargo & Company (SEC EDGAR): Form 10-K 2025, main document with annual report (wfc-20251231.htm) · 24 February 2026 · primary source
- Banking Dive: Wells Fargo won't face class action in mortgage discrimination lawsuit · 7 August 2025 · secondary source
2 · Safety
In its 10-K Wells Fargo documents a concrete cyber and third-party programme overseen by the Risk Committee, but AI-specific testing and red-teaming only via trade press.
What is documented
- Item 1C describes incident response, cyber and technology risk committees, internal and external reviews and annual training4
- A third-party programme with notification duties is overseen by the Risk Committee, to which the CISO and Head of Technology report4 5
- The annual report names hallucinations and data protection as GenAI risks with ongoing validation1
- The 10-Q for the third quarter of 2026 explicitly links AI to cyber risks6
- According to trade press, Fargo tokenises personal data before the external language model and filters before and after the model7
The case against
AI testing only from secondary sources
Sources for Safety (10)
- Wells Fargo & Company (SEC EDGAR, Form ARS): 2025 Annual Report (wfc2025annualreport-arsfil.pdf) · 3 November 2026 · primary source
- American Banker: How Wells Fargo is deploying the White House's AI Bill of Rights · 17 February 2023 · secondary source
- arXiv (via alphaXiv): Model Risk Management for Generative AI In Financial Institutions (arXiv:2503.15668) · 19 March 2025 · secondary source
- Wells Fargo & Company (SEC EDGAR, XBRL): Form 10-K 2025, R37 "Cybersecurity Risk Management and Strategy Disclosure" · 24 February 2026 · primary source
- Wells Fargo & Company (SEC EDGAR): Proxy Statement 2026 (Form DEF 14A, wfc-20260318.htm) · 18 March 2026 · primary source
- Wells Fargo & Company (SEC EDGAR): Form 10-Q for the third quarter of 2026 (wfc-20260930.htm) · 3 November 2026 · primary source
- VentureBeat: Wells Fargo's AI assistant just crossed 245 million interactions with zero humans in the loop and zero PII to the LLM · 8 April 2025 · secondary source
- The Daily Upside: Wells Fargo Patent Highlights Fatal Flaw in Deepfake Detection Models · 14 August 2025 · secondary source
- JD Supra: Wells Fargo Announces Data Breach Involving Unauthorized Access by Former Employee · 19 September 2024 · secondary source
- Crowdfund Insider: CFPB Drops Zelle Fraud Lawsuit Against Major US Banking Institutions · 4 March 2025 · secondary source
3 · Transparency
Wells Fargo discloses its AI use mainly in the shareholder letter and interviews, but the Fargo page does not mention AI and model or system information is missing.
What is documented
The case against
What is not disclosed
- The customer-facing Fargo page does not mention AI; labelling of automated offers is documented only in a 2023 interview2 3
- No model or system cards, AI transparency report or training data information were found8
- The good ranking in the Evident index does not document specific disclosures9
- The limits of the AI systems are not explained to customers
Sources for Transparency (9)
- Wells Fargo & Company (SEC EDGAR, Form ARS): 2025 Annual Report (wfc2025annualreport-arsfil.pdf) · 3 November 2026 · primary source
- Wells Fargo Stories: Get to know Wells Fargo's virtual assistant, Fargo · 3 November 2026 · secondary source
- American Banker: How Wells Fargo is deploying the White House's AI Bill of Rights · 17 February 2023 · secondary source
- Wells Fargo Stories: Wells Fargo, artificial intelligence, and you · 3 November 2026 · secondary source
- VentureBeat: Wells Fargo's AI assistant just crossed 245 million interactions with zero humans in the loop and zero PII to the LLM · 8 April 2025 · secondary source
- Wells Fargo Stories: How Wells Fargo builds responsible artificial intelligence · 22 September 2026 · secondary source
- United States District Court, Northern District of California (via Orrick InfoBytes): In re Wells Fargo Mortgage Discrimination Litigation, No. 22-cv-00990-JD, Order re Class Certification (Dkt. 328) · 5 August 2025 · primary source
- Wells Fargo & Company: Sustainability Disclosure Index (August 2026) · 3 November 2026 · secondary source
- Evident Insights: Evident AI Index for Banks, 2026 Edition · 3 November 2026 · secondary source
4 · Why the harm rule does not apply
The harm rule does not apply: the mortgage lawsuit is pending without a finding and concerns rules and scorecards that, according to Wells Fargo, use no AI; the CFPB lawsuit on Zelle was withdrawn.
5 · Controversies (3)
- mediumAccording to a lawsuit, automated mortgage processes disadvantage applicants; in 2025 the court denied class certification without finding discrimination. · 5 August 2025 · open
- minorThe CFPB sued over inadequate fraud protection at Zelle and withdrew the lawsuit with prejudice in 2025. · 4 March 2025 · abgewiesen
- minorA former employee accessed customer data without authorisation and partly used it for fraud; Wells Fargo reported the case. · 19 September 2024 · unklar
6 · What remains open
- The Risk Management, Risk Factors and Legal Actions chapters of the 2025 annual report, which could not be read
- Wording of an AI policy or AI principles and their scope
- Approval and review procedures for purchased AI models from external providers
- Customer-facing labelling of AI and complaints channels for AI-supported decisions
- Status of individual lawsuits after the denied class certification on mortgage lending
7 · How it was assessed
Cut-off date: 3 November 2026 · 40 searches, 44 fetches.
Primary documents read (10)
- Wells Fargo & Company Form 10-K 2025 (Part III and references; R37 Cybersecurity)
- Wells Fargo 2025 Annual Report (shareholder letter and Financial Review to p. 19)
- Proxy Statement 2026 (to p. 34, with further reading)
- Form 8-K Voting Results 2024, 2025 and 2026
- Form 10-Q Q3 2026 (to Credit Risk Management)
- Order re Class Certification, In re Wells Fargo Mortgage Discrimination Litigation (N.D. Cal., 2025-08-05)
- SEC Press Release 2025-16
- Sustainability Disclosure Index (August 2026)
- Wells Fargo newsroom releases on OCC, CFPB and Fed orders
- Wells Fargo Stories on Responsible AI and Fargo
Research of 3 November 2026 under Globalance AI Footprint rulebook version 0