Skip to content
GLOBALANCE INTELLIGENCEThe economy ahead.
This Week Index 28/100 Sign in
Money & Investments › GI 100 Index
The Great Value ShiftInvestment ThemesGI 100 IndexGI PortfolioStart Future Check
AI FOOTPRINTlimitedASSESSMENT 2026GLOBALANCEINTELLIGENCE
Globalance AI Footprint

Assessment as of November 2026 · next assessment 2027.

Result of the assessment by the Globalance research team under the published methodology. Not a certification or quality label.

How we assess ›

How this rating is builtVideo · 0:52
GI 100 Index · Profile

Morgan Stanley

USAAdoptersAssignment FullData quality Medium

Key messagelimited AI Footprint

Morgan Stanley uses AI mainly as an aid for advisers and specialists and has clearly regulated oversight of technology and cybersecurity at board level. The bank has not published binding rules on how it tests and approves its AI tools before use, and discloses which data it uses for them only in general terms.

Disclosure: the research was supported by an AI model made by Anthropic. Morgan Stanley is connected to Anthropic (e.g. as investor, partner, supplier, customer or through people). The assessment was therefore reviewed by people.

Opportunities

  • Safetysolid

    For technical system and data protection, the 10-K documents a cybersecurity programme aligned with the CRI profile (NIST CSF), penetration tests, red-team exercises, an incident response plan, third-party reviews and an annual independent third-party review reported to the O&T Committee; …

    For technical system and data protection, the 10-K documents a cybersecurity programme aligned with the CRI profile (NIST CSF), penetration tests, red-team exercises, an incident response plan, third-party reviews and an annual independent third-party review reported to the O&T Committee; generative AI is considered as a new threat (S02, B). For reliability and testing of the AI itself there are only company and vendor statements on evals before deployment, daily regression tests and validation by Model Risk Management (S13, S14, S17, C). For identity protection, VoiceID with monitoring by the fraud department is documented (S19, B), while safeguards for the planned agents with money movements are described only in interviews (S33, S34, C). The 2023 data protection settlement with six states over insecurely disposed storage devices shows a data protection control weakness without an AI connection (S20, A).

Risks

  • Responsibilitylimited

    AI-specific oversight is documented at grade B: the charter of the Operations and Technology Committee (as of 2026-10-18) requires the committee to receive management reports on technology strategy including AI and to approve material technology policies, and the 2026 proxy statement documents …

    AI-specific oversight is documented at grade B: the charter of the Operations and Technology Committee (as of 2026-10-18) requires the committee to receive management reports on technology strategy including AI and to approve material technology policies, and the 2026 proxy statement documents in-depth board sessions on AI (S06, S04); a Head of Firmwide AI and an AI steering group are documented only through a press release and the trade press (S12, S15, S16, C). There is no company-published AI policy with operational content and no described approval or risk classification procedure backed by grade A or B evidence: the risk-based approval procedure for GenAI applications and independent validation before go-live can only be inferred from job postings and an interview (S17, S18, S14, C), and AI figures or an AI report are missing. Under the minimum requirement for “solid”, the pillar is therefore “limited” on the basis of principles, without a documented material weakness.

  • Transparencylimited

    Morgan Stanley discloses AI purposes and the use of personal data to train its own models in its privacy policies only in general terms and promises notice where the law requires it (S07, S08, C).

    Morgan Stanley discloses AI purposes and the use of personal data to train its own models in its privacy policies only in general terms and promises notice where the law requires it (S07, S08, C). Press releases name OpenAI as the provider, consent for Debrief, source references for AskResearchGPT and general notes on possible errors and bias (S11, S12, C). Model or system cards, a transparency report, labelling rules towards clients or information on the agents were not found; only an approach is therefore documented.

  • Risk area

    AI agents with authority over money movements and account changes

  • Risk area

    AI-supported advisory materials with a risk of wrong recommendations to investors

  • No serious controversy recorded.

What lies behind the assessment

Research of 3 November 2026
43searches
37pages and documents read
11primary documents
3controversies checked
6open points, named openly
The full research: pillars, evidence, controversies and open points
AI Footprint · Adopter · As of 3 November 2026
Morgan StanleyOverall rating: limited

Morgan Stanley regulates oversight of technology and cybersecurity at board level and has its security programme reviewed externally every year. Approval and testing of its AI tools are documented only through press releases, job postings and interviews.

Responsibility

40 %
limitedsolidstrong

A board committee deals with AI, but documented AI approval rules are missing.

  • +O&T Committee charter names AI and requires reports on it Source ↗
  • +Board holds in-depth sessions on AI Source ↗
  • +Head of Firmwide AI since 2024 reporting to the co-presidents Source ↗
  • –Approval and validation only inferable from job postings Source ↗
  • –No published AI policy and no AI figures
All sources for this pillar (15)

Safety

40 %
limitedsolidstrong

Well-documented cybersecurity programme, but AI testing only as company and vendor statements.

  • +Cybersecurity programme with annual independent third-party review Source ↗
  • +VoiceID voice authentication monitored by the fraud department Source ↗
  • –AI testing documented only as company and vendor statements Source ↗
  • –USD 6.5 million settlement over disposed client data Source ↗
All sources for this pillar (12)

Transparency

20 %
limitedsolidstrong

AI purposes disclosed only in general terms, no labelling for clients and no system information.

  • +Debrief only with client consent Source ↗
  • +AskResearchGPT links to its sources Source ↗
  • –AI purposes and training on personal data disclosed only in general terms Source ↗
  • –No labelling of AI content for clients found
  • –No system descriptions of the tools and agents
All sources for this pillar (6)
What is missing for “strong”

These steps are in the company’s own hands.

1Responsibility
Officially document AI governance

Publish an AI policy with risk classification and approval procedure and explicitly set out board responsibility for AI and model risks.

2Safety
Document AI testing and agent safeguards

Disclose test results of the AI tools and safeguards for agents with money movements.

3Transparency
Label AI for clients

Label AI-generated content and agent interactions and describe the systems in use with their limitations.

Context: 101 companies assessed

Distribution of overall ratings, no ranking
strong
none yet
solid
limited
← Morgan Stanley is one of 66

No company reaches “strong” today. Disclosure on how companies handle AI is still young across the market.

Controversies
3 3 minor

No deduction for controversies.

Largest risk area
AI agents with access to client funds

Planned agents are to handle client queries, account changes and money movements, but their safeguards are described only in interviews.

How thoroughly assessed
33 Evidence

11 primary documents read, 43 searches.

Read the full reasoning and all sources
Morgan Stanley · Full reasoning

Board oversight of technology, but no published AI approval rules

Overall rating “limited”. No deduction for controversies. The superscript numbers link directly to the source.

1 · Responsibility

limitedWeight 40%

The Operations and Technology Committee charter names AI, but approval procedures and validation can only be inferred from job postings and an interview, and no AI policy is published.

What is documented
  • The O&T Committee receives reports on technology strategy including AI, approves material technology policies and tracks supervisory findings1
  • The 2026 proxy statement documents in-depth board sessions on AI2
  • Since March 2024 there has been a Head of Firmwide AI and a firm-wide AI steering group4 5 6
  • According to job postings, there is a risk-based approval framework for GenAI applications and independent validation by Model Risk Management7 8
  • Staff who develop or use AI receive bias training11
The case against
Weak evidence base
  • Approval procedures and validation before go-live can only be inferred from job postings and an interview8 7 9
  • The charter names AI only as a trend; there is no explicit board responsibility for AI and model risks1
Missing accountability
  • The wording of an AI policy, a risk classification and rules for purchased AI are not public
  • The 2024 and 2025 sustainability reports contain no AI governance figures; no external audit of AI governance was found14 15
Sources for Responsibility (15)
  1. Morgan Stanley: Operations and Technology Committee Charter · 18 October 2026 · primary source
  2. Morgan Stanley / U.S. SEC: Proxy Statement 2026 (Form DEF 14A) · 2 April 2026 · primary source
  3. Morgan Stanley / U.S. SEC: Proxy Statement 2025 (Form DEF 14A) · 4 April 2025 · primary source
  4. Banking Dive: Morgan Stanley names Jeff McMillan head of firmwide AI · 15 March 2024 · secondary source
  5. Financial Planning: Morgan Stanley gives AI even bigger role with new appointment · 15 March 2024 · secondary source
  6. Morgan Stanley: Morgan Stanley Research Announces AskResearchGPT · 23 October 2024 · secondary source
  7. AnitaB.org (Stelleninserat Morgan Stanley): AI Governance and Compliance Specialist (Hybrid) (job posting) · 3 November 2026 · secondary source
  8. AnitaB.org (Stelleninserat Morgan Stanley): Generative AI – Model Risk Management Associate (job posting) · 3 November 2026 · secondary source
  9. The American Bazaar: Jeff McMillan on Morgan Stanley's human-first AI governance · 26 January 2026 · secondary source
  10. Morgan Stanley / U.S. SEC: Form 10-K for the fiscal year ended 2025-12-31 · 31 December 2025 · primary source
  11. Morgan Stanley: Artificial Intelligence and Machine Learning (Privacy and AI Statement, U.S. Privacy Policy and Notice) · 25 April 2025 · secondary source
  12. Morgan Stanley: Morgan Stanley Wealth Management Announces Latest Game-Changing Addition to Suite of GenAI Tools (AI @ Morgan Stanley Debrief) · 26 June 2024 · secondary source
  13. Morgan Stanley: EMEA Privacy Policy · 1 April 2025 · secondary source
  14. Morgan Stanley: Morgan Stanley 2025 Sustainability Report · 31 December 2025 · secondary source
  15. Morgan Stanley: Morgan Stanley 2024 Sustainability Report · 31 December 2024 · secondary source

2 · Safety

solidWeight 40%

The 10-K documents a cybersecurity programme with an annual independent third-party review, but testing of the AI tools and agents rests on company and vendor statements.

What is documented
  • The cybersecurity programme follows the CRI profile (NIST CSF) with penetration tests, red-team exercises, an incident response plan and third-party reviews5
  • An annual independent third-party review is reported to the O&T Committee; generative AI is considered as a new threat5 6
  • VoiceID stores no recordings; the fraud department reviews risks10
  • According to a vendor case study, there are evals for each use case before deployment and daily regression tests; no data is retained at OpenAI1
The case against
AI testing weakly documented
  • Test results, error rates or external audits of the AI systems are not published1 2 3
  • Safeguards for the planned agents with money movements are described only in interviews12 11
Data protection weakness without AI connection
  • Six states received USD 6.5 million because storage devices with unencrypted client data were disposed of without adequate oversight7 8
Sources for Safety (12)
  1. OpenAI: Morgan Stanley (customer case study) · 3 November 2026 · secondary source
  2. The American Bazaar: Jeff McMillan on Morgan Stanley's human-first AI governance · 26 January 2026 · secondary source
  3. AnitaB.org (Stelleninserat Morgan Stanley): Generative AI – Model Risk Management Associate (job posting) · 3 November 2026 · secondary source
  4. Morgan Stanley / U.S. SEC: Form 10-K for the fiscal year ended 2025-12-31 · 31 December 2025 · primary source
  5. Morgan Stanley / U.S. SEC (XBRL): Form 10-K 2025, R34: Cybersecurity Risk Management and Strategy Disclosure · 31 December 2025 · primary source
  6. Morgan Stanley: Operations and Technology Committee Charter · 18 October 2026 · primary source
  7. New York State Office of the Attorney General: Attorney General James and Multistate Coalition Secure $6.5 Million from Morgan Stanley · 16 November 2023 · primary source
  8. Compliance Week: Morgan Stanley settles with states for $6.5m over mishandled data · 17 November 2023 · secondary source
  9. The Star (Reuters/NYT): Major banks including JPMorgan and Citi warned of data exposure after hack, NYT reports · 23 November 2025 · secondary source
  10. Morgan Stanley: VoiceID Client Guide · 1 September 2023 · primary source
  11. Financial Planning: Behind the Marvel-movie inspired AI 'super agents' coming to Morgan Stanley · 11 February 2026 · secondary source
  12. Banking Dive / CIO Dive: Morgan Stanley's wealth division AI playbook · 26 October 2026 · secondary source

3 · Transparency

limitedWeight 20%

Morgan Stanley states AI purposes and data use in its privacy policies only in general terms, and information on the agents and on labelling for clients is missing.

What is documented
  • The Privacy and AI Statement lists AI purposes and promises notice where there is a material effect1
  • The EMEA privacy policy mentions AI and machine learning, automated decisions and training of its own models with personal data2
  • Debrief creates meeting notes only with consent; users are alerted to possible errors and bias3
  • AskResearchGPT names GPT-4 as its base model and links to the sources4
The case against
Missing labelling and system information
  • No rule on labelling AI-generated content or agent interactions for clients was found
  • There are no explanations of recommendations from machine learning («next-best action») or of the limitations of the agents6
  • Information on data sources, retention and providers of the agents and a transparency report are missing
Sources for Transparency (6)
  1. Morgan Stanley: Artificial Intelligence and Machine Learning (Privacy and AI Statement, U.S. Privacy Policy and Notice) · 25 April 2025 · secondary source
  2. Morgan Stanley: EMEA Privacy Policy · 1 April 2025 · secondary source
  3. Morgan Stanley: Morgan Stanley Wealth Management Announces Latest Game-Changing Addition to Suite of GenAI Tools (AI @ Morgan Stanley Debrief) · 26 June 2024 · secondary source
  4. Morgan Stanley: Morgan Stanley Research Announces AskResearchGPT · 23 October 2024 · secondary source
  5. OpenAI: Morgan Stanley (customer case study) · 3 November 2026 · secondary source
  6. Banking Dive / CIO Dive: Morgan Stanley's wealth division AI playbook · 26 October 2026 · secondary source

4 · Why the harm rule does not apply

The harm rule does not apply: the fines and settlements in the period have no AI connection, and no pattern of serious incidents involving its own AI systems was found.

5 · Controversies (3)

6 · What remains open

  • Public wording of an AI policy, risk classification and approval procedure
  • Explicit board responsibility for AI and model risks
  • Test and monitoring results of the AI tools and agents
  • Safeguards for agents with money movements and for access by external agents
  • Labelling of AI-generated content for clients
  • Status of the anti-money-laundering investigations and the legal proceedings in the 10-K, which could not be read

7 · How it was assessed

Cut-off date: 3 November 2026 · 43 searches, 37 fetches.

Primary documents read (11)
  • Morgan Stanley Form 10-K 2025 (Part I to Risk Factors; XBRL R22 and R34)
  • Morgan Stanley Form 10-Q as of 2026-06-30 (Part I to Liquidity)
  • Proxy Statement 2026 (complete) and Proxy Statement 2025 (first 100,000 characters)
  • Operations and Technology Committee Charter (as of 2026-10-18)
  • Morgan Stanley 2025 and 2024 Sustainability Reports (partial)
  • Privacy and AI Statement (USA) and EMEA Privacy Policy
  • VoiceID Client Guide
  • Pressemitteilungen zu AI @ Morgan Stanley Debrief und AskResearchGPT
  • Press release by the NY Attorney General on the 2023 settlement
  • SEC Press Release 2024-193
  • SEC no-action filing relating to the shareholder proposal by the NYC Retirement Systems

Research of 3 November 2026 under Globalance AI Footprint rulebook version 0

We assess documented practice in dealing with AI; this is not an investment recommendation. Three pillars, weights 40/40/20, harm rule.

Research of 3 November 2026 under Globalance AI Footprint rulebook version 0. Documented observations from public sources, no investment recommendation.

Important Notice. This material is provided for information purposes only and does not constitute an offer or solicitation to buy or sell any financial instrument, nor does it constitute investment advice or a recommendation. The GI 100 Index is a rules-based strategy index of Globalance Bank. Investments involve risk, including the loss of capital invested. The final product documents (including FinSA) shall prevail in all cases.

Your company? Information on inclusion in the GI 100 Index →

What comes next