How does HSBC govern AI responsibly, including at HSBC USA?
HSBC governs AI through group-wide principles and committees that also cover HSBC USA. Its Principles for the Ethical Use of Data and AI, in a version dated July 2024, set seven commitments, starting with a rule that HSBC does not use AI decision-making 'for a purpose that could not be justified if pursued through human decision-making'. Novel AI use cases are reviewed and approved by an AI Review Committee of senior executives, fed by an AI Centre of Excellence in a hub-and-spoke model; in the first half of 2025 the group embedded AI Review Councils across the organization. The Annual Report and Accounts 2025 says models using AI or GenAI are reviewed by risk teams and that HSBC has enhanced inventory control 'to apply heightened scrutiny of agentic AI use cases before deployment'. In the United States, HSBC USA Inc.'s 10-K filings describe a Model Risk Committee chaired by the chief risk officer and a Model Risk Management function that approves models used in HSBC North America, and say the bank trains leaders and model developers on the ethical-use principles. The 10-K risk factors moved from 'generative artificial intelligence' in the filings for 2023 and 2024 to 'artificial intelligence' in the filing for 2025, which adds that the rising use of AI escalates the risk of cyberattacks. HSBC does not publish a US AI use-case inventory, AI incident data or a separate US AI policy.[1][2][3][4][5][6][7]
Data as of Oct 31, 2026 · HSBC profile updated Oct 19, 2026 · 16 sources cited on this page · 141 banks profiled on this site · Markdown
What bodies and documents govern AI at HSBC?
| Element | What it does | Since / as of | Source |
|---|---|---|---|
| Principles for the Ethical Use of Data and AI | Seven principles, from 'Act consistently with HSBC's values' to 'Contribute to development of best practice'; includes 'We establish clear accountability for our AI systems' | Latest version July 2024 | HSBC[1] |
| AI Review Committee | A collection of senior executives that reviews and approves novel AI use cases; ideas come from business lines ('spokes') through a central hub that includes the AI Centre of Excellence | Described by co-chair Steve Albrecht | HSBC[2] |
| AI Review Councils | Embedded across the organization during 1H25, building on the Group AI Review Committee, with AI lifecycle management refined | 1H25 | HSBC AI page[3] |
| Agentic AI controls | Models using AI or GenAI reviewed by risk teams; inventory control enhanced 'to apply heightened scrutiny of agentic AI use cases before deployment' | Annual Report 2025 | HSBC[4] |
| Model Risk Committee (HSBC North America) | Chaired by the CRO, with representation from businesses and functions; the Model Risk Management function provides effective challenge to and approval of models | 10-K for 2024 and 2025 | SEC EDGAR[5][6] |
| Training | HSBC USA supports 'leaders and model developers with training that helps them understand and apply our principles on the ethical use of big data and artificial intelligence'; the group runs mandatory training covering responsible AI | 10-K for 2023 to 2025; AI page 2026 | SEC EDGAR; HSBC[7][6][3] |
| Human in the loop | Pilots run with limited numbers of users before scaled production; 'we need to have a "human in the loop" and other controls to help us spot errors' | Described by Steve Albrecht | HSBC[2] |
How do HSBC USA's SEC filings describe AI risk?
HSBC USA Inc. 10-K risk factors and model-risk discussion, filing by filing.
| Filing | What it says about AI |
|---|---|
| 10-K for 2023 (filed February 21, 2024) | 'The accessibility and increasing sophistication of generative artificial intelligence brings financial crime risks'; criminals may use it 'to perpetrate fraud, particularly scams'[7] |
| 10-K for 2024 (filed February 19, 2025) | 'Artificial intelligence and machine learning models remain a key focus', with governance 'particular focus on generative artificial intelligence'; more sophisticated AI modelling 'could also lead to increased model related risks'[5] |
| 10-K for 2025 (filed February 25, 2026) | Wording moves from generative AI to 'artificial intelligence' throughout; adds that 'the rising use of artificial intelligence also escalate the overall risk of cyberattacks'[6] |
| Group Annual Report 2025 | Regulation is 'increasingly fragmented, including in relation to AI'; focus remains on AI and machine learning models, 'including the development of GenAI and agentic AI'[4] |
Where does HSBC take part in setting AI standards, and what does it not disclose?
HSBC is one of eight firms behind the July 2026 white paper on runtime safeguards for agentic finance, and on July 20, 2026 it became founding member of the Emerging Payments Association Asia's AI & Agentic Payments Working Group, which addresses who is liable when an AI agent exceeds its mandate and how agents are identified across borders.[8][9]
The 2025 annual report lists 'publishing HSBC's Principles for the Ethical Use of Data and AI' among its mitigating actions for new technology risk, and says the group has governance in place 'on AI and digital assets and currencies'.[4]
Not disclosed: the members of the AI Review Committee, how many use cases it has rejected, AI-related incidents or complaints, bias-testing results, and any AI governance body specific to HSBC USA. The 10-K filings describe North America model risk governance but do not count AI models separately.[6][2]
What does the HSBC profile record on this topic?
The rows below are the entries of the HSBC AI strategy profile that bear on this page, as of Oct 19, 2026; the profile holds the full record of 34 dated moves, 16 use cases and 53 sources.
| Use case | What the record shows | Status | Source |
|---|---|---|---|
| AI in the US model-risk and risk-factor framework | The 10-K says more sophisticated AI modelling could increase model risk, for which capital is held; a CRO-chaired Model Risk Committee manages model risk and the Model Risk Management function approves models used in HSBC North America. | In production | [6] |
| AI in financial-crime risk management | Group programme with Google Cloud to detect risk earlier and intervene twice as fast across nearly one billion transactions monitored a month; the US 10-K notes AI's potential for financial-crime detection and its use by criminals. | Rolling out | [10][6] |
| Agentic payments (Mastercard) | Partner bank for AI-agent-initiated payments; co-author of runtime safeguards for agentic finance. | Pilot | [11][8] |
| Dynamic Risk Assessment (Google Cloud AML AI) | HSBC co-developed the system with Google, piloted it in 2021 and says it screens about 980 million transactions a month with 60% fewer false positives; the June 2026 Google deal targets agentic AI for earlier intervention. | In production | [12][13] |
| Date | Move | Detail | Source |
|---|---|---|---|
| Oct 9, 2026 | Ant Digital test: AI agents pay with HSBC tokenised deposits | A technical verification test, not a launch, used the Tokenised Deposit Service, now available in the US, with real-time risk checks through HSBC's MCP interface. | [14] |
| Aug 4, 2026 | HSBC USA 10-Q names generative AI as a forward-looking risk | Development and use of AI, including generative AI, by HUSI, group affiliates, vendors, customers and competitors, with operational, cyber, fraud and regulatory risks. | [15] |
| Jul 3, 2026 | SAFR: safeguards for agentic finance at runtime | HSBC is one of eight firms behind the white paper with JPMorganChase, Mastercard and others. | [8] |
| Feb 25, 2026 | HSBC USA 10-K flags AI model, fraud and cyber risk | The US holding company's annual report cites AI-driven model risk, AI-enabled scams and AI-escalated cyberattacks among its risk factors. | [6] |
| Oct 21, 2025 | Marrying AI with accountability in payments | A look at how HSBC frames AI in global money movement: data, regulation and risk as one choreography. | [16] |
| Jun 10, 2024 | HSBC describes Dynamic Risk Assessment, built with Google | HSBC says the AI financial-crime system it co-developed with Google, piloted in 2021, checks about 980 million transactions a month, finds two to four times more financial crime and cuts false positives by 60%. | [12] |
Does HSBC have published AI principles?
Yes. HSBC's Principles for the Ethical Use of Data and AI, latest version July 2024, set seven principles: act consistently with HSBC's values, protect privacy, start with a clearly defined purpose, address unfair bias, be responsible for AI, adapt governance to emerging needs, and contribute to best practice. HSBC USA's 10-K says leaders and model developers are trained on them.
Who approves new AI use cases at HSBC?
Novel AI use cases are reviewed and approved by HSBC's AI Review Committee, a group of senior executives, after business lines submit ideas to a central hub that includes the AI Centre of Excellence. In the first half of 2025 the group added AI Review Councils across the organization. Pilots then run with limited users before scaled production.
How does HSBC USA manage AI model risk?
Through HSBC North America's model risk framework: a Model Risk Committee chaired by the chief risk officer, a Model Risk Policy, and a Model Risk Management function that challenges and approves models, per the 10-K. The 10-K for 2024 says AI and machine learning models 'remain a key focus', with particular attention to generative AI.
What does HSBC do about agentic AI risk?
The 2025 annual report says HSBC enhanced inventory control to apply heightened scrutiny of agentic AI use cases before deployment. HSBC also co-authored a July 2026 white paper on runtime safeguards for agentic finance and is a founding member of an Asia-Pacific working group on standards for agentic payments.
Has HSBC USA reported any AI incidents or enforcement?
None is disclosed. HSBC USA's 10-K filings describe AI as a source of model, fraud and cyber risk but report no AI-specific incident, and the group publishes no AI incident or complaint data. BankingNewsAI's AI incidents page tracks public cases across banks.
- HSBC's Principles for the Ethical Use of Data and AI (latest version: July 2024)
- 'There's a risk in missing the AI opportunity'
- Transforming HSBC with AI (content last updated 17 August 2026)
- HSBC Holdings plc Annual Report and Accounts 2025
- HSBC USA Inc. Annual Report on Form 10-K for the year ended December 31, 2024
- HSBC USA Inc. Annual Report on Form 10-K for the year ended December 31, 2025
- HSBC USA Inc. Annual Report on Form 10-K for the year ended December 31, 2023
- Your Bank's AI Agent May Need a Permission Slip
- HSBC joins EPAA agentic AI working group as founding member
- HSBC and Google Cloud Announce Transformative AI Banking Partnership
- US card networks accelerate bets on agentic AI
- Harnessing the power of AI to fight financial crime
- HSBC and Google Cloud announce transformative AI banking partnership
- Ant Digital Technologies and HSBC announce successful AI agent micropayment technical verification test
- HSBC USA Inc. Quarterly Report on Form 10-Q for the quarter ended June 30, 2026
- Inside HSBC's Push to Marry AI With Accountability
When HSBC 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 PT · free
plus every regulation, bank and agent page update, the morning after · leave any morning