How does Regions govern AI responsibly?
Regions governs AI through existing model-risk machinery plus a generative-AI layer added after ChatGPT. Manav Misra, chief data and analytics officer, told Citi's FinTech conference in November 2025 that machine learning had run for six to seven years with guardrails across all three lines of defense, and that generative AI brought a new risk framework, a responsible use policy and a standard for unstructured data so that sensitive data does not train external models. At board level, the 2026 proxy says the Technology Committee, chaired by Zhanna Golodryga, monitored management's deployment of generative AI, its foundational controls and priority use cases 'to help ensure AI is introduced in a responsible, risk-aligned manner', and the Risk Committee tracked model performance including generative AI. The 2026 investor decks set three guiding principles for ATLAS, the gen-AI gateway: 'Human in The Loop Design', 'Managed Risk' and 'Secure AI Platform'. The 2025 10-K carries a stand-alone AI risk factor covering incorrect or biased output, opacity, third-party models and evolving AI law. Regions Bank is on NIST's February 2024 member list of the US AI Safety Institute Consortium. It has not published a standalone AI principles document or any bias-testing results.[1][2][3][4][5][6]
Data as of Oct 16, 2026 · Regions profile updated Oct 13, 2026 · 15 sources cited on this page · 140 banks profiled on this site · Markdown
Which bodies, policies and principles govern AI at Regions?
What Regions has stated or filed, with the body responsible. 'Stated' means said by an executive; 'Filed' means in an SEC document.
| Item | What it says | Status | Date |
|---|---|---|---|
| Model risk management across three lines of defense | Traditional machine-learning models evaluated independently for six to seven years so they do not introduce additional risk | Stated (Misra) | November 19, 2025[1] |
| Generative-AI risk framework and responsible use policy | Added after ChatGPT: a new risk framework, new policies including a responsible use policy, new standards including one for unstructured data; aim that 'none of our sensitive data left the bank and was used to train external models' | Stated (Misra) | November 19, 2025[1] |
| Federated data governance | Business units own their data, report issues and fix them; enterprise data management covering governance, lineage and quality 'so that we can have a level of confidence in the data that feeds these models' | Stated | November 19, 2025[1] |
| Board Technology Committee | Monitored generative-AI deployment, foundational controls and priority use cases; reviewed milestones, timelines and safeguards; chaired by Zhanna Golodryga; five meetings in 2025 | Filed (2026 proxy) | March 23, 2026[2] |
| Board Risk Committee | Monitored key risks including 'model performance (including generative AI)' and third-party risk management | Filed (2026 proxy) | March 23, 2026[2] |
| Director skills and training | 'Business Operations and Technology Innovation/AI' is a board skill; director education covers 'cyber and information security (including AI)' | Filed (2026 proxy) | March 23, 2026[2] |
| Generative-AI guiding principles | 'Human in The Loop Design', 'Managed Risk', 'Secure AI Platform'; use cases built on shared patterns that 'help avoid fragmentation' | Investor deck | February and May 2026[3][4] |
| US AI Safety Institute Consortium (NIST) | Regions Bank listed among consortium members | Member | February 7, 2024[6] |
| Industry standards work | Chris Kennedy, SVP and head of enterprise payments, says he has contributed to responsible-AI standards through the Data & Trust Alliance | Stated (speaker profile) | 2026[7] |
| Human oversight for business clients | Commercial bankers: 'We want intelligent automation doing paperwork, not relationship work'; the bank tells clients to ask 'What human oversight is involved in important decisions?' | Published | September 14, 2026[8] |
What does Regions' 10-K say about AI risk?
From 'The development and use of AI presents risks and challenges that may adversely impact our business' in the 2025 Form 10-K (filed February 24, 2026). The 2026 10-Qs repeat the heading without new detail.
| Risk | What the 10-K says |
|---|---|
| Regulation | The legal and regulatory environment for AI is 'uncertain and rapidly evolving' and could require changes to how Regions implements AI and raise compliance costs[5] |
| Output quality and bias | AI models, 'particularly generative AI models, may, and often do, produce output or take action that is incorrect', release confidential information or reflect discrimination, errors or biases[5] |
| Explainability | Model complexity makes it hard to understand outputs, complicating compliance with laws 'that require documentation or explanation of the basis on which decisions are made'[5] |
| Third-party models | 'We currently rely, and expect to continue to rely, on AI models developed by third parties', with risks from their training data and possible disclosure of Regions' information into public training sets[5] |
| Filings since | The Q1 and Q2 2026 10-Qs list the same AI risk-factor heading[9][10] |
What has Regions not published about responsible AI?
Regions has no standalone AI principles document, responsible-AI report or named AI ethics officer on the public record. Its responsible-use policy and generative-AI risk framework were described at an investor conference, not published. The three ATLAS principles appear only as a slide in the investor decks.[1][3]
No fairness or bias-testing results have been disclosed for the machine-learning models the 10-K says Regions uses in pricing, loan grading, fraud prevention and authentication, nor for the personalised insights launched to mobile customers in July 2026. The bank's public writing on AI risk has focused on customers: AI-assisted scams, deepfakes and voice cloning.[5][11][12][13]
What does the Regions profile record on this topic?
The rows below are the entries of the Regions AI strategy profile that bear on this page, as of Oct 13, 2026; the profile holds the full record of 33 dated moves, 15 use cases and 44 sources.
| Use case | What the record shows | Status | Source |
|---|---|---|---|
| AI and machine-learning model risk (10-K) | Stand-alone AI risk factor plus machine-learning models used in pricing, loan grading, fraud prevention, authentication and lead prospecting. | In production | [5] |
| Traditional and generative AI across the company | Expanded on data governance, authentication and real-time data foundations. | Rolling out | [14] |
| Deepfake and AI-scam education | Client guidance on AI-generated impersonation. | In production | [13] |
| Generative-AI risk framework and responsible use policy | Added after ChatGPT: a risk framework, a responsible use policy and a standard for unstructured data, on top of model-risk management across three lines of defence; sensitive data must not train external models. | In production | [1] |
| Date | Move | Detail | Source |
|---|---|---|---|
| Aug 6, 2026 | Q2 2026 10-Q filed | The risk-factor list again carries the AI heading; the Q2 call itself addresses AI only through credit exposure to the AI ecosystem. | [10][15] |
| May 7, 2026 | Q1 2026 10-Q repeats the AI risk-factor headline | The filing lists the 10-K heading on AI risks and challenges among its risk factors without new AI detail. | [9] |
| Feb 24, 2026 | 2025 10-K carries a stand-alone AI risk factor | Generative-AI errors, bias, data leakage and opacity, evolving AI law, and AI-enhanced cyberattacks; machine-learning models named across pricing, credit and fraud. | [5] |
| Nov 19, 2025 | Citi FinTech conference: modernisation and AI | Massey, Misra and Weiss describe a rebuilt API layer, a cloud data layer called Customer DNA, six to seven years of machine learning, a new generative-AI risk framework and $70 million of tech and operations savings. | [1] |
| Mar 11, 2025 | Regions publishes guidance on AI-assisted fraud | Head of Financial Crimes Prevention Sue Ross says fraudsters use large language models to write more convincing scams and voice cloning to impersonate relatives and bosses. | [12] |
Does Regions have a responsible AI policy?
Yes, but it is not published. Chief data and analytics officer Manav Misra told Citi's FinTech conference in November 2025 that after ChatGPT the bank added a new generative-AI risk framework, new policies including a responsible use policy, and a standard for unstructured data. Regions has not released the documents themselves.
Which board committee oversees AI at Regions?
The Technology Committee, chaired by Zhanna Golodryga, which met five times in 2025. The 2026 proxy says it monitored management's deployment of generative AI, its foundational controls and priority use cases. The Risk Committee separately monitors model performance including generative AI, and directors receive training on cyber and information security including AI.
What are Regions' generative AI principles?
The February and May 2026 investor decks list three guiding principles for the generative-AI strategy: 'Human in The Loop Design', 'Managed Risk' and 'Secure AI Platform'. Use cases are built on shared patterns, such as enterprise knowledge enablement and intelligent document analysis, which the deck says help avoid fragmentation and improve scalability and maintainability.
Is Regions part of any AI safety or standards body?
NIST's February 2024 member list of the US AI Safety Institute Consortium includes Regions Bank. Chris Kennedy, Regions' SVP and head of enterprise payments, says in his 2026 American Banker speaker profile that he has contributed to responsible-AI standards through the Data & Trust Alliance.
What AI risks does Regions disclose to investors?
The 2025 10-K has a stand-alone risk factor: evolving AI law, generative output that may be incorrect, leak confidential data or reflect bias, models too complex to explain, reliance on third-party AI models, and AI-related system failures and cyber-attacks. The 2026 10-Qs repeat the heading.
- Regions Financial at Citi's FinTech Conference: Modernization and AI Focus
- Regions Financial Corporation 2026 Proxy Statement (Schedule 14A)
- Regions Financial investor meeting materials, February 2026 (Form 8-K, Exhibit 99.1)
- Regions Financial investor meeting materials, May 2026 (Form 8-K, Exhibit 99.1)
- Regions Financial Corporation Annual Report on Form 10-K for 2025
- The US AI Safety Institute Consortium Member List
- Chris Kennedy, SVP, Head of Enterprise Payments, Regions Bank (Digital Banking 2026 speaker profile)
- AI in banking: What it can mean for your business
- Regions Financial Corporation Quarterly Report on Form 10-Q for the quarter ended 31 March 2026
- Regions Financial Corporation Quarterly Report on Form 10-Q for the quarter ended 30 June 2026
- Regions Bank Launches Personalized Insights for Mobile Banking Customers
- How to detect and avoid fraud in the age of AI
- The new face of fraud
- Regions Bank Says AI Lifts Productivity 20% as Loan Growth Cools
- Regions Financial Q2 Earnings Call Highlights
When Regions moves on AI, you'll read it here first.
when one of the banks we cover moves on AI, it is in the next morning's brief · six sourced stories · 7 am ET · free
plus every tracker, bank and agent page update, the morning after · leave any morning