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 FOOTPRINTsolidASSESSMENT 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

Pfizer Inc

USAAdoptersAssignment FullData quality Medium

Key messagesolid AI Footprint

Pfizer has principles for artificial intelligence, an internal group body for risk management and an information security programme audited by third parties, and its Health Answers health assistant openly shows sources, limits and unverified answers. Harm from Pfizer's own AI is not documented in the period; figures, external audits and information on its other AI applications are largely missing.

Opportunities

  • Safetysolid

    Item 1C of the 10-K for FY2025 documents a specific security programme (NIST alignment, regular third-party assessments of maturity and the control environment, internal and external penetration testing, supplier risks, periodically tested incident response plans, Audit Committee oversight) (S04, S05).

    Item 1C of the 10-K for FY2025 documents a specific security programme (NIST alignment, regular third-party assessments of maturity and the control environment, internal and external penetration testing, supplier risks, periodically tested incident response plans, Audit Committee oversight) (S04, S05). Specific to AI, Health Answers is documented with sample checks of AI answers by independent physicians (S17); for VOX and Charlie there are only company and partner accounts (S15, S16). No incident involving its own AI systems is documented in the period; pending are a class action over website tracking unrelated to AI and a data breach at the service provider Cencora (S21, S22, S23). Solid, because the grade B evidence mainly concerns the general cyber programme.

  • Transparencysolid

    For the patient-facing Health Answers application it is specifically documented that it is labelled as generative AI, links sources for each answer, Seals unverified answers as «unverified» and warns of hallucinations and limits; the privacy policy discloses the use of AI with personal data (S17, S19).

    For the patient-facing Health Answers application it is specifically documented that it is labelled as generative AI, links sources for each answer, Seals unverified answers as «unverified» and warns of hallucinations and limits; the privacy policy discloses the use of AI with personal data (S17, S19). Beyond that there are only principles to inform users «where practicable» about limits and risks (S01, S02). For the other applications (Charlie, VOX, research models, manufacturing) there are no labelling rules, system or model information or an AI inventory. Only narrowly solid and with need for review, because the grade B evidence covers a single product.

Risks

  • Responsibilitylimited

    The 2025 and 2026 proxy statements document that the board explicitly treats AI as a risk topic and that the Audit Committee oversees technology risks (S06, S07); the 2024 Impact Report names an AI risk management programme under a cross-functional AI Council with a policy, training, risk review …

    The 2025 and 2026 proxy statements document that the board explicitly treats AI as a risk topic and that the Audit Committee oversees technology risks (S06, S07); the 2024 Impact Report names an AI risk management programme under a cross-functional AI Council with a policy, training, risk review and controls (S11), co-chaired by the Chief AI and Analytics Officer (S13). However, a published AI policy with operational content or a described approval, risk classification or review procedure is missing: the Impact Report only lists the programme's components, the code of conduct generally requires “appropriate measures to assess and mitigate risks” and human oversight without naming responsibilities, procedures or prohibitions (S11, S12), and the group policy itself is not published. Under the minimum requirement for “solid”, the pillar is therefore “limited”, on the basis of principles and without a documented material weakness; reporting channels with investigation by Compliance are documented (S12).

  • Risk area

    Patient-facing generative AI providing health information to lay people

  • Risk area

    AI in drug safety, clinical development and regulatory documents

  • Risk area

    AI in manufacturing and supply chain affecting drug quality

  • No serious controversy recorded.

What lies behind the assessment

Research of 3 November 2026
41searches
35pages and documents read
10primary documents
2controversies checked
6open points, named openly
The full research: pillars, evidence, controversies and open points
AI Footprint · Adopter · As of 3 November 2026
Pfizer IncOverall rating: solid

Pfizer has documented AI principles, an AI risk programme under an AI Council and a security programme audited by third parties. Specific labelling and notices on limits exist only for the Health Answers health assistant.

Safety

40 %
limitedsolidstrong

The security programme is audited by third parties; AI testing is documented almost only for Health Answers.

  • +Regular third-party assessments and penetration tests per 10-K Source ↗
  • +Independent physicians check samples of AI answers Source ↗
  • –Item 1C of the 10-K does not mention AI Source ↗
  • –No validation results or red-teaming reports published
  • –Class action over website tracking despite opt-out, unrelated to AI Source ↗
All sources for this pillar (13)
What is missing for “strong”

These steps are in the company’s own hands.

1Responsibility
Publish an operational AI policy

Publish the AI policy with its approval procedure, assign AI to a board committee and have the programme externally audited with figures.

2Safety
Disclose testing of its own AI

Publish validation results, error rates and red-teaming for Health Answers and other in-house AI systems.

3Transparency
Labelling beyond Health Answers

Disclose labelling rules for AI advertising content, an AI inventory and the models used.

Context: 101 companies assessed

Distribution of overall ratings, no ranking
strong
none yet
solid
← Pfizer Inc is one of 35
limited

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

Controversies
2 2 minor

No deduction for controversies.

Largest risk area
Misinformation from patient-facing AI

Health Answers answers lay people's health questions with generative AI; Pfizer itself warns that inaccurate or incomplete answers may occur.

How thoroughly assessed
31 Evidence

10 primary documents read, 41 searches.

Read the full reasoning and all sources
Pfizer Inc · Full reasoning

AI risk programme and audited IT security, but evidence almost only for Health Answers

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

1 · Responsibility

limitedWeight 40%

Pfizer documents board oversight of AI risks and an AI risk programme under an AI Council, but publishes neither the policy nor an approval procedure.

What is documented
  • The board's risk oversight names «cybersecurity and AI»; the Audit Committee oversees information security and technology risks1 2
  • An AI risk management programme with principles, a group policy, training, risk review and controls sits under a cross-functional AI Council3
  • The AI Council is co-chaired by the Chief AI and Analytics Officer, in office since 20244
  • The code of conduct requires measures to assess and mitigate risks when developing, procuring and deploying AI6
  • A Compliance Helpline is available around the clock and anonymously, with investigation by Compliance and a ban on retaliation6
The case against
No published operational AI rules
  • The group AI policy, the approval process and a risk classification are not published
  • There are no AI metrics such as an inventory, reviewed applications or training rate and no external audit or certification
Unclear responsibility
  • Neither the charter nor the proxy explicitly assigns AI to a board committee; the mandate and current status of the AI Council are not disclosed1
  • The description of the AI Council comes from the 2024 Impact Report because the AI section of the 2025 report could not be read3 10
Sources for Responsibility (11)
  1. Pfizer Inc. / U.S. Securities and Exchange Commission: Proxy Statement 2026 (Form DEF 14A) · 12 March 2026 · primary source
  2. Pfizer Inc. / U.S. Securities and Exchange Commission: Proxy Statement 2025 (Form DEF 14A) · 13 March 2025 · primary source
  3. Pfizer Inc.: Pfizer 2024 Impact Report · 31 December 2024 · primary source
  4. pharmaphorum: Pfizer grows digital team with chief AI officer appointment · 18 August 2024 · secondary source
  5. American Healthcare Leader: Pfizer AI (profile on AI compliance and generative AI) · 30 September 2024 · secondary source
  6. Pfizer Inc.: 2025 Blue Book: Pfizer's Code of Conduct · 1 January 2025 · primary source
  7. Pfizer Inc. (Global Policy & Public Affairs): Policy Position on Artificial Intelligence · 11 December 2023 · secondary source
  8. Pfizer Inc. (pfizer.com): Three Principles of Responsibility for Artificial Intelligence (AI) in Healthcare · 3 November 2026 · secondary source
  9. Digiday: With 'Charlie,' Pfizer is building a new generative AI platform for pharma marketing · 22 February 2024 · secondary source
  10. Pfizer Inc.: Pfizer 2025 Impact Report · 31 December 2025 · secondary source
  11. Pfizer Inc.: Human Rights Policy Statement (version of 28 May 2026) · 28 May 2026 · secondary source

2 · Safety

solidWeight 40%

Pfizer describes in its 10-K a specific security programme audited by third parties, but AI-specific testing is documented only for Health Answers.

What is documented
  • Item 1C describes NIST alignment, regular third-party assessments of maturity and the control environment, and internal and external penetration testing5 6
  • Supplier risks, periodically tested incident response plans and Audit Committee oversight are documented5
  • For Health Answers, independent physicians check articles and samples of AI answers against clinical guidelines; an independent advisory board reviews the guidelines1
  • The Health Answers privacy policy excludes sale and sharing for behavioural advertising8
The case against
Missing evidence of AI testing
  • Validation results, red-teaming and error rates of its own AI systems, including Health Answers, are not published
  • For the VOX and Charlie platforms there are only company and partner accounts2 3
  • AI-specific protections against prompt injection or data leakage to AI providers are not described
Data incidents unrelated to AI
  • According to a class action, pfizer.com passed data to Google trackers despite «Decline All»; the proceedings are ongoing9 10
  • In 2024 data from Pfizer patient programmes leaked at the service provider Cencora11 12
Sources for Safety (13)
  1. Pfizer Inc. (healthanswers.pfizer.com): Health Answers by Pfizer – How it Works · 3 November 2026 · primary source
  2. Digiday: With 'Charlie,' Pfizer is building a new generative AI platform for pharma marketing · 22 February 2024 · secondary source
  3. Amazon Web Services (AWS for Industries Blog): Three takeaways from Pfizer's AWS re:Invent keynote · 28 November 2023 · secondary source
  4. Pfizer Inc. (pfizer.com): AI in Drug Safety: Building the Elusive 'Loch Ness Monster' of Reporting Tools · 3 November 2026 · secondary source
  5. Pfizer Inc. / U.S. Securities and Exchange Commission: Form 10-K FY2025, XBRL R29: Cybersecurity Risk Management and Strategy Disclosure · 26 February 2026 · primary source
  6. Board Cybersecurity (board-cybersecurity.com): PFIZER INC 10-K Item 1C. Cybersecurity – 2026-02-26 · 26 February 2026 · secondary source
  7. Pfizer Inc.: Pfizer 2024 Impact Report · 31 December 2024 · primary source
  8. Pfizer Inc. (healthanswers.pfizer.com): Health Answers by Pfizer – Protecting your privacy (Privacy Policy) · 1 January 2025 · primary source
  9. Top Class Actions: Pfizer class action alleges website tracked users despite cookie opt-outs · 22 May 2026 · secondary source
  10. U.S. District Court S.D.N.Y. (via Leagle): Nakamura v. Pfizer Inc., Nos. 26-CV-3914 (JMF), 26-CV-5683 (JMF), Order · 23 July 2026 · primary source
  11. The HIPAA Journal: Cencora Cyberattack and Data Breach · 21 February 2024 · secondary source
  12. CyberMaterial: Pfizer Data Breach Affects Personal Info · 7 June 2024 · secondary source
  13. Board Cybersecurity (board-cybersecurity.com): PFIZER INC Cybersecurity Incident (Tracker) · 7 June 2024 · secondary source

3 · Transparency

solidWeight 20%

For Health Answers Pfizer discloses the use of AI, sources, unverified answers and limits; for Charlie, VOX and research models, labelling and model information are missing.

What is documented
  • Health Answers is labelled «powered by generative AI» and links the sources for each answer1 2
  • Answers without independent sources are labelled «unverified»; Pfizer warns of hallucinations and incomplete answers1
  • The Health Answers privacy policy discloses the use of AI with personal data3
  • The source corpus of Health Answers is named, for example Harvard Health Publishing, CDC, FDA and NIH1
The case against
Only one product documented
  • For Charlie, VOX, research models and manufacturing there are no labelling rules, system information or an AI inventory
  • According to its principles Pfizer informs about limits and risks only «where practicable»5 6
  • The language model and provider behind Health Answers are not disclosed1
Sources for Transparency (7)
  1. Pfizer Inc. (healthanswers.pfizer.com): Health Answers by Pfizer – How it Works · 3 November 2026 · primary source
  2. Pfizer Inc. (pfizer.com): Introducing Health Answers by Pfizer: New Consumer Digital Product Providing Answers · 3 November 2026 · secondary source
  3. Pfizer Inc. (healthanswers.pfizer.com): Health Answers by Pfizer – Protecting your privacy (Privacy Policy) · 1 January 2025 · primary source
  4. Pfizer Inc. / U.S. Securities and Exchange Commission: Form 10-K for the fiscal year ended December 31, 2025 · 26 February 2026 · primary source
  5. Pfizer Inc. (Global Policy & Public Affairs): Policy Position on Artificial Intelligence · 11 December 2023 · secondary source
  6. Pfizer Inc. (pfizer.com): Three Principles of Responsibility for Artificial Intelligence (AI) in Healthcare · 3 November 2026 · secondary source
  7. Pharmaceutical Technology (pharmaceutical-tech.com): Chai Discovery Signs Licensing Agreement with Pfizer to Advance AI-Driven Drug Discovery · 5 June 2026 · secondary source

4 · Why the harm rule does not apply

The harm rule does not apply: no AI-related harm, decision or settlement is documented within the window; the tracking lawsuit and the Cencora incident are unrelated to AI.

5 · Controversies (2)

6 · What remains open

  • Wording of the AI risk factor in the FY2025 10-K, which was truncated on retrieval
  • Current status and mandate of the AI Council and AI reporting to the board
  • AI-specific metrics, validation results and external audits
  • Model used and test results of Health Answers
  • Labelling and review rules for AI-generated advertising content on the Charlie platform
  • AI section «Responsible Technology» of the 2025 Impact Report

7 · How it was assessed

Cut-off date: 3 November 2026 · 41 searches, 35 fetches.

Primary documents read (10)
  • Form 10-K FY2025 (cover page, forward-looking information, Item 1 incl. Data Privacy)
  • Form 10-K FY2025 XBRL R29 (Item 1C Cybersecurity) and R25 (Contingencies, in part)
  • Proxy Statements 2026, 2025 and 2024, and DEFA14A 2025
  • Pfizer 2024 Impact Report (pp. 16, 23, 26, 27) and 2025 Impact Report (in part)
  • Policy Position on Artificial Intelligence (December 2023)
  • 2025 Blue Book (Code of Conduct)
  • Human Rights Policy Statement (May 2026)
  • Health Answers by Pfizer: How it Works and Privacy Policy
  • DOJ press release on the Biohaven settlement (24 January 2025)
  • Court order, Nakamura v. Pfizer (23 July 2026, via Leagle)

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