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

Amazon

USAInnovatorAssignment SegmentData quality Medium

Key messagesolid AI Footprint

Amazon has published traceable testing, security and labelling procedures for its own AI models and cloud services and has part of them externally certified. At the same time, lawsuits are pending over facial recognition in Ring cameras and in virtual try-on, and in France a fine for storing data on warehouse workers for too long was upheld in court.

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

Opportunities

  • Responsibilitysolid

    Board oversight of AI is documented in the 2026 proxy statement: the Nominating and Corporate Governance Committee oversees risks from “responsible AI development and AI governance” and reviewed the approach during the reporting year; cybersecurity lies with the Security Committee, and the Chief …

    Board oversight of AI is documented in the 2026 proxy statement: the Nominating and Corporate Governance Committee oversees risks from “responsible AI development and AI governance” and reviewed the approach during the reporting year; cybersecurity lies with the Security Committee, and the Chief Security Officer reports to the CEO (S1, S3, S4). For its own frontier models, the Frontier Model Safety Framework documents thresholds, escalation paths and go/no-go decisions, and four AWS services hold an ISO/IEC 42001 certification with a surveillance audit without findings (S9, S13, S14). Coverage, however, is limited to AWS and the Nova models: Ranking Digital Rights rates governance at 19% and human rights due diligence at 0% (S21), a review of customer use of AI and cloud services by military and government customers is not disclosed (S40, S41), and proposals on AI oversight were rejected or not put to a vote in 2026 (S5, S42).

  • Safetysolid

    For its own Nova models, testing is documented concretely: more than 300 internal attack techniques, external red-teaming by ActiveFence, Deloitte, CMU and Nemesys, frontier testing with METR and independent comparative tests (S10, S11, S12); the FMSF describes the protection of model weights, and …

    For its own Nova models, testing is documented concretely: more than 300 internal attack techniques, external red-teaming by ActiveFence, Deloitte, CMU and Nemesys, frontier testing with METR and independent comparative tests (S10, S11, S12); the FMSF describes the protection of model weights, and Bedrock has documented abuse detection with CSAM reporting (S9, S16). Against this stand documented weaknesses: in 2025 a deletion command made its way into a shipped version of the AI assistant Amazon Q via an overly broad GitHub token (CVE-2025-8217, fixed; S26), media report outages linked to AI tools (S27, S28), and the Conseil d'État upheld a GDPR infringement in the retention of performance data in French warehouses (S24). Class actions without findings are pending over Ring's facial recognition and virtual try-on (S33, S35); there is no established serious AI harm, which is why the pillar narrowly remains “solid”.

  • Transparencysolid

    For Nova, Amazon publishes a technical report with a model card, AWS AI Service Cards with limits of use and a report under the G7 Hiroshima framework (S11, S12, S14); image and video outputs are labelled with invisible watermarks and in part with C2PA metadata, with a detection tool (S12, S18).

    For Nova, Amazon publishes a technical report with a model card, AWS AI Service Cards with limits of use and a report under the G7 Hiroshima framework (S11, S12, S14); image and video outputs are labelled with invisible watermarks and in part with C2PA metadata, with a detection tool (S12, S18). Training data is described only by category, and a lawsuit accuses Amazon of using YouTube videos for Nova Reel by circumventing protective measures (S11, S37). Ranking Digital Rights rates disclosure on data processing and algorithmic systems as weak (S21), and Ring does not disclose the results of its accuracy tests (S31).

Risks

  • Risk area

    Facial recognition in Ring cameras and in virtual try-on

  • Risk area

    AI and cloud services for military and security agencies

  • Risk area

    AI development agents with access to code and production environments

  • Risk area

    Algorithmic performance monitoring of logistics workers

  • No serious controversy recorded.

What lies behind the assessment

Research of 3 November 2026
42searches
47pages and documents read
17primary documents
10controversies checked
5open points, named openly
The full research: pillars, evidence, controversies and open points
AI Footprint · Innovator · As of 3 November 2026
AmazonOverall rating: solid

Amazon documents concrete testing, security and labelling procedures for its Nova models and AWS services and has some of them externally certified. Ring, Alexa and logistics are not covered, and class actions over facial recognition are pending.

Responsibility

40 %
limitedsolidstrong

A committee oversees AI, but the rules mainly cover AWS and the Nova models.

  • +A board committee oversees AI governance Source ↗
  • +Frontier framework with thresholds and a go/no-go decision Source ↗
  • +ISO 42001 certification for four AWS services Source ↗
  • –No disclosed due diligence for government and military customers Source ↗
  • –Ranking Digital Rights: human rights due diligence is missing Source ↗
All sources for this pillar (15)

Safety

40 %
limitedsolidstrong

The Nova models are externally tested, but there were incidents and biometrics lawsuits are pending.

  • +External red-teaming of the Nova models by several third parties Source ↗
  • +Abuse detection in Bedrock with CSAM reporting Source ↗
  • –A deletion command reached a shipped version of Amazon Q Source ↗
  • –Conseil d'État upheld a privacy breach in French warehouses Source ↗
  • –Class actions over Ring Familiar Faces and virtual try-on Source ↗
All sources for this pillar (21)

Transparency

20 %
limitedsolidstrong

Model cards and watermarks are in place, but information on training data remains general.

  • +Technical report with model card and service cards for Nova Source ↗
  • +Invisible watermarks and C2PA for generated images and videos Source ↗
  • –Training data described only by category Source ↗
  • –Lawsuit alleges Nova Reel was trained on YouTube videos Source ↗
  • –Ring does not disclose the results of its accuracy tests Source ↗
All sources for this pillar (8)

AI resource use

Whole group incl. data centres · reporting year 2025
Disclosure partialDetailsClose
9.4 million m³ water100 % renewable80.8 million t CO₂eNet zero by 2040, emissions +16% on prior year
Electricity
–not disclosed
Water withdrawal
9.4 million m³
Renewable electricity
100 %
Total CO₂e
80.8 million t
Target
by
Status 2025
Net-zero carbon across operations (Climate Pledge)
2040
80.9m t CO₂e, +16% on prior year
Match 100% of electricity with renewables (was 2030)
2025
100% (2025, third year running)
Water positive across data centres
2030
75% of goal (2024: 53%)

Targets and status as reported by the company. Electricity is matched on an annual basis; AWS may buy renewable attributes from other regions.

Amazon emitted a total of 80.85 million t CO2e in 2025; operations and electricity (Scopes 1+2) came to 37.23 million t location-based and 19.15 million t market-based. Its data centers withdrew 9.4 million m³ of water at a WUE of 0.12 L/kWh; Amazon does not disclose its electricity consumption.

  • 2026 Employees criticise rising emissions and the pace of AI in an open letter
  • 2026 Data centre water use disclosed for the first time after transparency criticism
Sources

Supplement without a score: not part of the Globalance AI Footprint.

What is missing for “strong”

These steps are in the company’s own hands.

1Responsibility
Extend AI rules group-wide

Extend the framework and certification to Ring, Alexa and logistics and disclose due diligence for government and military customers.

2Safety
Investigate and document incidents

Publish root cause analyses for Amazon Q and Kiro and the remediation following the decision in France.

3Transparency
Disclose training data and test results

Describe training data sources and rights clearance concretely and publish Ring's accuracy tests.

Context: 101 companies assessed

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

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

Controversies
10 6 medium, 4 minor

No deduction for controversies.

Largest risk area
Facial recognition without third-party consent

According to a class action, Ring's Familiar Faces feature captures facial data of visitors and passers-by without their consent.

How thoroughly assessed
44 Evidence

17 primary documents read, 42 searches.

Read the full reasoning and all sources
Amazon · Full reasoning

Tested in-house models and cloud services, pending lawsuits over facial recognition

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

1 · Responsibility

solidWeight 40%

Amazon has board oversight of AI and a safety framework for frontier models, but discloses no due diligence for government and military customers.

What is documented
  • The Nominating and Corporate Governance Committee oversees risks from responsible AI development and AI governance2 3
  • A Security Committee of independent directors is responsible for security; the Chief Security Officer reports to the CEO1
  • The Frontier Model Safety Framework defines thresholds for CBRN, cyber and automated AI research and provides for a go/no-go decision by senior leadership4
  • Bedrock, Q Business, Textract and Transcribe are certified to ISO/IEC 42001; the 2025 surveillance audit found nothing10
  • A central Responsible AI Governance Team is described in the report under the G7 Hiroshima framework5
The case against
Limited scope
  • The certification and framework apply to AWS services and Nova models, not to Ring, Alexa or logistics10 4
  • Ranking Digital Rights rates human rights due diligence at 0% and governance at 19%7
  • A group-wide AI policy with risk classes is not disclosed
Government customers and shareholders
  • Investigations accuse Amazon of accepting clauses in the Project Nimbus contract with Israel that undermine its own terms of use8
  • No review of the use of AI and cloud services by military and security agencies is disclosed9
  • Two AI proposals were rejected in 2025; a proposal on the Responsible AI Approach was not put to a vote in 202613 15
Sources for Responsibility (15)
  1. Amazon.com, Inc. (SEC-Einreichung): Form 10-K for fiscal year 2025 · 3 November 2026 · primary source
  2. Amazon.com, Inc.: 2026 Proxy Statement (PDF) · 3 November 2026 · primary source
  3. Amazon.com, Inc. (SEC Filing): DEF 14A 2026 · 3 November 2026 · primary source
  4. Amazon.com, Inc. (Amazon Science): Amazon's Frontier Model Safety Framework · 3 November 2026 · primary source
  5. OECD.AI (Report by Amazon): G7 Hiroshima AI Process Reporting Framework – Amazon (Nova models on Bedrock) · 31 October 2025 · primary source
  6. Amazon Web Services: AWS Responsible AI Policy · 13 January 2025 · secondary source
  7. Ranking Digital Rights: 2025 Big Tech Edition – Amazon · 13 May 2025 · primary source
  8. Democracy Now!: Report: Google and Amazon violated their own terms of service in Israel's $1.2 billion Project Nimbus deal · 31 October 2025 · secondary source
  9. Intercommunity Peace & Justice Center (Aktionärsantrag): Shareholder proposal on aligning AI sales and use with the Responsible AI Approach · 3 November 2026 · secondary source
  10. Amazon Web Services (AWS Security Blog): AWS successfully completed its first surveillance audit for ISO 42001:2023 with no findings · 26 February 2026 · primary source
  11. Amazon.com, Inc. (About Amazon): Responsible AI – Our approach · 3 November 2026 · secondary source
  12. Business & Human Rights Resource Centre: US corporate human rights index – Amazon (findings table) · 1 August 2025 · secondary source
  13. Amazon.com, Inc. (SEC-Einreichung): Form 8-K, voting results of the 2025 annual general meeting · 21 May 2025 · primary source
  14. Interfaith Center on Corporate Responsibility (ICCR): Amazon 2025 Shareholder Proposals · 23 May 2025 · secondary source
  15. Investor Advocates for Social Justice (SEC PX14A6G): Notice of Exempt Solicitation for the 2026 annual general meeting · 3 November 2026 · secondary source

2 · Safety

solidWeight 40%

Amazon documents extensive internal and external testing of its Nova models, but a compromised AI assistant, a confirmed privacy breach and biometrics lawsuits stand against this.

What is documented
  • The Nova technical report cites more than 300 internal attack techniques and external red-teaming by ActiveFence, Deloitte, CMU and Nemesys3
  • The service card documents frontier testing with METR; PRISM Eval and ActiveFence tested Nova Premier independently2 1
  • Model weights are protected with encryption and isolation; vulnerabilities can be reported via a bug bounty6
  • Bedrock detects abuse automatically, reports CSAM to NCMEC and can suspend accounts11
  • For Familiar Faces, Ring cites opt-in, deletion periods and internal privacy impact assessments14
The case against
Incidents with in-house AI tools
  • A deletion command reached version 1.84.0 of the Amazon Q extension via an overly broad GitHub token; AWS has fixed the flaw8 9
  • Media report an AWS outage after changes made by the agent tool Kiro; Amazon cites user error4
  • Amazon Quick lacked an authorisation check, so blocked AI agents were reachable; the gap was fixed10
Privacy and biometrics
  • The Conseil d'État upheld a breach of data minimisation in performance data of warehouse workers and reduced the fine to EUR 15 million18
  • According to the lawsuit, Ring Familiar Faces captures facial data of visitors and passers-by without consent; the case is pending16
  • A class action under the Illinois biometrics law has been certified over virtual try-on; liability has not been decided17
Sources for Safety (21)
  1. Amazon.com, Inc. (Amazon Science): Independent evaluations demonstrate Nova Premier's safety · 29 May 2025 · primary source
  2. Amazon Web Services: AWS AI Service Card: Amazon Nova Micro, Lite, Pro, and Premier · 3 November 2026 · primary source
  3. Amazon AGI (arXiv): The Amazon Nova Family of Models: Technical Report and Model Card · 17 March 2025 · primary source
  4. The Register: Amazon's vibe-coding tool Kiro reportedly vibed too hard and brought down AWS · 20 February 2026 · secondary source
  5. Fortune: Amazon retail site outages and AI agent inaccurate advice · 12 March 2026 · secondary source
  6. Amazon.com, Inc. (Amazon Science): Amazon's Frontier Model Safety Framework · 3 November 2026 · primary source
  7. Amazon.com, Inc. (SEC-Einreichung): Form 10-K for fiscal year 2025 · 3 November 2026 · primary source
  8. Amazon Web Services: Security Bulletin AWS-2025-015: Amazon Q Developer Extension for Visual Studio Code (Version #1.84) · 23 July 2025 · primary source
  9. BleepingComputer: Amazon AI coding agent hacked to inject data wiping commands · 24 July 2025 · secondary source
  10. Pulse (via Fog Security): Amazon Quick authorisation bypass let users reach blocked AI chat agents · 12 May 2026 · secondary source
  11. Amazon Web Services: Amazon Bedrock abuse detection (user guide) · 3 November 2026 · primary source
  12. Amazon Web Services: AWS Responsible AI Policy · 13 January 2025 · secondary source
  13. ConductAtlas (via AWS Service Terms): Amazon Rekognition police department moratorium · 3 November 2026 · secondary source
  14. Ring LLC (Brief an Senator Markey, veröffentlicht durch das Büro Markey): Ring's reply to the letter of 30 October 2025 on Familiar Faces · 21 November 2025 · primary source
  15. Electronic Frontier Foundation: The Legal Case Against Ring's Face Recognition Feature · 3 November 2025 · secondary source
  16. U.S. District Court, Western District of Washington (CourtListener): Sigwalt v. Amazon.com, Inc. and Ring LLC, No. 2:26-cv-01887, complaint · 1 June 2026 · secondary source
  17. U.S. Court of Appeals for the Seventh Circuit: Svoboda v. Amazon.com Inc., No. 25-1361 (class certification ruling) · 17 December 2025 · primary source
  18. Conseil d'État (France): Décision n° 492830, Amazon France Logistique · 23 December 2025 · primary source
  19. Actu-Juridique: La CNIL sanctionne Amazon pour utilisation d'un système de surveillance excessivement intrusif · 23 January 2024 · secondary source
  20. The Register: Amazon to kill off local Alexa processing, all voice requests shipped to the cloud · 17 March 2025 · secondary source
  21. Tom's Hardware: Amazon's Rufus AI shopping assistant can be easily jailbroken · 9 March 2026 · secondary source

3 · Transparency

solidWeight 20%

Amazon publishes model cards and service cards and Seals generated images and videos, but describes training data only by category.

What is documented
  • For the Nova models there is a technical report with a model card and AWS AI Service Cards stating limits of use1 5
  • Amazon has published a report under the G7 Hiroshima framework on the Nova models2
  • Outputs of Nova Canvas, Reel and Sonic carry invisible watermarks; there is a detection tool for Titan images1 3
  • The service cards recommend human review for consequential decisions and disclosure to end users5
  • Amazon has signed the EU code of practice for general-purpose AI models4
The case against
Training data
  • Training data is described only as licensed, proprietary, open-source and publicly available data5 1
  • According to the lawsuit, Amazon used millions of YouTube videos for Nova Reel by circumventing protective measures; Amazon did not comment7
Disclosure for consumer products
  • Ranking Digital Rights rates disclosure on algorithmic systems as weak6
  • Ring tests the accuracy of Familiar Faces by population group but does not disclose the results8
  • Labelling of AI-generated content in Alexa+ and Rufus is not documented
Sources for Transparency (8)
  1. Amazon AGI (arXiv): The Amazon Nova Family of Models: Technical Report and Model Card · 17 March 2025 · primary source
  2. OECD.AI (Report by Amazon): G7 Hiroshima AI Process Reporting Framework – Amazon (Nova models on Bedrock) · 31 October 2025 · primary source
  3. Amazon Web Services: Watermark detection for Amazon Titan Image Generator now available in Amazon Bedrock · 23 April 2024 · primary source
  4. Euronews: Some 26 tech companies sign up to EU Commission's AI code · 1 August 2025 · secondary source
  5. Amazon Web Services: AWS AI Service Card: Amazon Nova Micro, Lite, Pro, and Premier · 3 November 2026 · primary source
  6. Ranking Digital Rights: 2025 Big Tech Edition – Amazon · 13 May 2025 · primary source
  7. GeekWire: Amazon sued by YouTubers for allegedly scraping their content to train AI video tool · 3 November 2026 · secondary source
  8. Ring LLC (Brief an Senator Markey, veröffentlicht durch das Büro Markey): Ring's reply to the letter of 30 October 2025 on Familiar Faces · 21 November 2025 · primary source

4 · Why the harm rule does not apply

The harm rule does not apply: the AI link of the fine upheld in France is not established, and the other cases are pending proceedings without findings.

5 · Controversies (10)

Further controversies (4)

6 · What remains open

  • Full list of AI-related legal proceedings; Note 7 of the 10-K and the Q3 2026 10-Q could not be read
  • Remedial measures following the decision of the CNIL and the Conseil d'État
  • Disclosed human rights review of AI sales to government agencies and the military
  • Group-wide figures on AI incidents and complaints
  • No stand-alone responsible AI report from Amazon was found

7 · How it was assessed

Cut-off date: 3 November 2026 · 42 searches, 47 fetches.

Primary documents read (17)
  • Form 10-K FY 2025 (S1, characters 0–117,795)
  • Form 10-Q Q2 2026 (S2, characters 0–100,000)
  • 2026 Proxy Statement PDF (S3, characters 0–132,531)
  • DEF 14A 2026 (S4, characters 0–100,000)
  • Form 8-K Voting Results 2025 (S5)
  • Frontier Model Safety Framework (S9)
  • AWS AI Service Card Nova (S11)
  • Nova Technical Report and Model Card (S12)
  • HAIP Report Amazon (S14)
  • AWS Responsible AI Policy (S15)
  • Bedrock abuse detection (S16)
  • Security Bulletin AWS-2025-015 (S26)
  • Ring Letter to Senator Markey (S31)
  • Complaint Sigwalt v. Amazon (S33)
  • Judgment Svoboda v. Amazon, 7th Cir. (S35)
  • Conseil d'État 492830 (S24)
  • 2025 Amazon Sustainability Report (S20, characters 0–125,567)

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