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

Palantir Technologies

USAInnovatorAssignment CoreData quality High

Key messagelimited AI Footprint

Palantir describes principles on privacy and human rights and holds a government security authorisation for its cloud, but discloses little about how it vets the use of its software by the military, police and immigration authorities. Several organisations and investors criticise these deployments, and reports link a Palantir system to a deadly strike on a school in Iran, which the company denies.

Disclosure: the research was supported by an AI model made by Anthropic. Palantir Technologies 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 there is an official record (FedRAMP High, S26) and the 10-K disclosures on the CISO, board oversight, external audits and the absence of material cyber incidents (B, S11).

    For technical system and data protection there is an official record (FedRAMP High, S26) and the 10-K disclosures on the CISO, board oversight, external audits and the absence of material cyber incidents (B, S11). For reliability and protection against misuse there are only company statements (C) on evaluation, audit-log and approval features (S12, S30, S14). Against this stand medium allegations that have not been officially established: according to media reports on an unpublished Pentagon review, overreliance on Maven contributed to the school strike in Minab; Palantir denies any fault in its software (S03, S04, S05); in addition, Medicaid data that had been passed on unlawfully ended up in ICE's Palantir environment (S23). Because there is no serious incident documented at grade A or B, no pillar cap applies; the “solid” rating rests essentially on sub-criterion b and needs review.

Risks

  • Responsibilitylimited

    The proxy statement documents (B) that no board committee is assigned oversight of AI, human rights or ethics and that the board explicitly rejects independent human rights assessments; the founders control up to just under 50% of the voting rights (S06, S07).

    The proxy statement documents (B) that no board committee is assigned oversight of AI, human rights or ethics and that the board explicitly rejects independent human rights assessments; the founders control up to just under 50% of the voting rights (S06, S07). A Privacy and Civil Liberties team, a human rights policy, grievance channels and customer vetting before contracts are signed are documented only by the company itself (C) (S10, S15, S14). Independent bodies (Amnesty, EFF, NYU Stern, NYC Comptroller) find inadequate human rights due diligence in the work for ICE and Israel; Palantir did not answer accountability questions specifically (S29, S16, S36, S28). There is no finding by a court or authority against Palantir; the rating rests on the governance weakness documented at grade B, not on a serious incident.

  • Transparencylimited

    Palantir discloses AI risks in general terms in the 10-K and publishes responses to enquiries from the UN, Amnesty and the EFF (S10, S17, S29, S14).

    Palantir discloses AI risks in general terms in the 10-K and publishes responses to enquiries from the UN, Amnesty and the EFF (S10, S17, S29, S14). For the riskiest systems (Maven, ImmigrationOS, ELITE) there are no system or model cards, no information on limitations or errors and no transparency report. In the proxy statement the board rejects disclosure of human rights assessments, citing confidentiality and cost (B, S06); the shareholder proposals on this failed (B, S07).

  • Risk area

    Military targeting with the Maven Smart System

  • Risk area

    Immigration enforcement with ImmigrationOS and ELITE for ICE

  • No serious controversy recorded.

What lies behind the assessment

Research of 3 November 2026
37searches
38pages and documents read
10primary documents
10controversies checked
5open points, named openly
The full research: pillars, evidence, controversies and open points
AI Footprint · Innovator · As of 3 November 2026
Palantir TechnologiesOverall rating: limited

Palantir documents the technical protection of its cloud with a government authorisation, but discloses little about how it vets deployments with the military, police and immigration authorities. No board committee is assigned oversight of AI or human rights.

Responsibility

40 %
limitedsolidstrong

No board oversight of AI, and human rights assessments are rejected.

  • +Human rights policy referring to the UN Guiding Principles Source ↗
  • +Named lead for privacy and civil liberties Source ↗
  • –No board committee for AI, human rights or ethics Source ↗
  • –Board rejects independent human rights assessments Source ↗
  • –Amnesty: inadequate human rights due diligence in work for ICE Source ↗
All sources for this pillar (10)

Safety

40 %
limitedsolidstrong

Officially certified system protection, but allegations over Maven's role in a strike.

  • +FedRAMP High authorisation for Palantir Federal Cloud Service Source ↗
  • +According to the 10-K, external audits and no material cyber incidents Source ↗
  • –Reports: overreliance on Maven contributed to the school strike Source ↗
  • –Unlawfully shared Medicaid data ended up with Palantir Source ↗
  • –No published test and validation results for Maven
All sources for this pillar (17)
What is missing for “strong”

These steps are in the company’s own hands.

1Responsibility
Anchor AI oversight in the board

Assign oversight of AI and human rights to a board committee and commission an independent human rights assessment of high-risk deployments.

2Safety
Make high-risk systems verifiable

Publish test and validation results for Maven, ImmigrationOS and ELITE and have them independently reviewed.

3Transparency
Publish system cards

Disclose system cards with deployment limits, error rates and data sources for the most sensitive systems.

Context: 101 companies assessed

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

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

Controversies
10 4 medium, 6 minor

No deduction for controversies.

Largest risk area
Errors in military targeting

According to media reports on an unpublished Pentagon review, a school in Minab came out of Maven as a target because of outdated data; Palantir denies any fault in its software.

How thoroughly assessed
36 Evidence

10 primary documents read, 37 searches.

Read the full reasoning and all sources
Palantir Technologies · Full reasoning

Certified cloud security, but little oversight of military and immigration deployments

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

1 · Responsibility

limitedWeight 40%

Policies and grievance channels are documented only by the company itself, and the board explicitly rejects independent human rights assessments.

What is documented
  • According to its own human rights policy, Palantir screens risks by sector, country and capability, reviews new use cases and handles misuse reactively5
  • The board cites a review process involving legal, compliance, privacy, security and finance, and customer vetting before contracts are signed1
  • As grievance channels, Palantir names an ethics hotline, a third-party whistleblower hotline and a dedicated address for human rights matters5
  • Courtney Bowman is named as responsible for privacy and civil liberties engineering2 3
The case against
Oversight and control
  • No board committee is assigned oversight of AI, human rights or ethics1
  • The founders control up to just under 50% of the voting rights; proposals for an independent report and a human rights impact assessment failed1 7
  • No risk classification or approval procedure has been published for high-risk deployments such as Maven, ImmigrationOS and ELITE3
Criticism from independent bodies
  • Amnesty concludes that Palantir has not met its responsibility owing to a lack of adequate human rights due diligence8
  • According to the EFF, Palantir did not answer its accountability questions on the ICE work; the NYC Comptroller is calling for an independent risk assessment9 4
Sources for Responsibility (10)
  1. Palantir Technologies Inc. / SEC EDGAR: Form DEF 14A – Proxy Statement 2026 · 24 April 2026 · primary source
  2. Electronic Frontier Foundation: Letter from Cindy Cohn to Courtney Bowman (Palantir) · 13 February 2026 · secondary source
  3. Palantir Technologies Inc. (veröffentlicht durch EFF): Reply from Courtney Bowman to the EFF · 6 March 2026 · secondary source
  4. New York City Comptroller: Letter to Palantir Technologies Requesting Third-Party Human Rights Risk Assessment · 4 February 2026 · secondary source
  5. Palantir Technologies: Palantir Human Rights Policy · 3 November 2026 · secondary source
  6. Palantir Technologies (product documentation): AI ethics and governance (Foundry/AIP Docs) · 3 November 2026 · secondary source
  7. Palantir Technologies Inc. / SEC EDGAR: Form 8-K – voting results of the 2026 annual meeting · June 2026 · primary source
  8. Amnesty International: Palantir: Response from Palantir to Amnesty International (AMR 51/0211/2025) · August 2025 · secondary source
  9. Electronic Frontier Foundation: Palantir Has a Human Rights Policy. Its ICE Work Tells a Different Story · 20 April 2026 · secondary source
  10. NYU Stern Center for Business and Human Rights: As Palantir's Role in Government Grows, So Does the Need for Real Human Rights Due Diligence · 3 June 2025 · secondary source

2 · Safety

solidWeight 40%

Technical protection is officially certified, but there are only company statements on reliability and protection against misuse.

What is documented
  • Palantir Federal Cloud Service is certified to FedRAMP High, with six operating authorisations6
  • According to the 10-K, a CISO reports periodically to the board, independent auditors carry out external reviews, and cyber incidents have so far had no material impact7
  • According to the product documentation, the platform offers testing, versioning, rollback and human approvals1
  • According to the company, customer data is not used to train third-party models and is not stored with model providers8
The case against
Allegations about high-risk deployments
  • According to media reports on an unpublished Pentagon review, overreliance on Maven contributed to the school strike in Minab; Palantir denies any fault in its software3 2 4
  • According to court filings, unlawfully shared Medicaid data reached ICE's Palantir environment; according to Palantir, it was deleted15
  • According to the EFF, a user manual for ELITE disables filters so that all targets are displayed16
Missing evidence of testing
  • No validation or red-teaming results have been published for Maven, ELITE and ImmigrationOS
  • According to the media, an internal Swiss Army report stated that data leakage to US authorities could not be technically prevented9 10
Sources for Safety (17)
  1. Palantir Technologies (product documentation): AI ethics and governance (Foundry/AIP Docs) · 3 November 2026 · secondary source
  2. Defense News: Deadly Iran school strike casts shadow over Pentagon's AI targeting push · 24 March 2026 · secondary source
  3. Gizmodo (via Bloomberg): Pentagon Investigators Say Overreliance on Palantir AI Tech Contributed to U.S. Strike That Killed 123 Iranian Children · 20 September 2026 · secondary source
  4. AI Weekly (via Bloomberg): Pentagon Rewires Palantir Maven After Iran School Strike · 22 September 2026 · secondary source
  5. Benzinga (via Reuters): Palantir, Anduril say report of security flaws was outdated, inaccurate · 3 October 2025 · secondary source
  6. FedRAMP (US General Services Administration): FedRAMP Marketplace: Palantir Federal Cloud Service – Supporting Services (PFCS-SS) · 19 November 2024 · primary source
  7. Board Cybersecurity (verbatim from the 10-K): Palantir Technologies Inc. 10-K Item 1C. Cybersecurity – 2026-02-17 · 17 February 2026 · primary source
  8. Palantir Technologies (product documentation): AIP security and privacy · 3 November 2026 · secondary source
  9. netzpolitik.org (nach Republik/WAV): Switzerland: Palantir software carries devastating risks · 8 December 2025 · secondary source
  10. netzpolitik.org: Switzerland: “Die Republik” prevails against Palantir · 16 June 2026 · secondary source
  11. PiunikaWeb: Palantir hacked? CTO Shyam Sankar rebuts Kim Dotcom claim · 17 February 2026 · secondary source
  12. Palantir Technologies Inc. (veröffentlicht durch EFF): Reply from Courtney Bowman to the EFF · 6 March 2026 · secondary source
  13. Palantir Technologies (blog): Correcting the Record: Response to the EFF January 15, 2026 Report on Palantir · 3 November 2026 · secondary source
  14. Palantir Technologies Inc. / SEC EDGAR: Form 10-K for fiscal year 2025 · 17 February 2026 · primary source
  15. NPR: ICE shared Medicaid data it wasn't supposed to have with Palantir · 17 July 2026 · secondary source
  16. Electronic Frontier Foundation: Letter from Cindy Cohn to Courtney Bowman (Palantir) · 13 February 2026 · secondary source
  17. beck-aktuell: GFF files constitutional complaint against Bavaria's rules for Palantir police software · 24 July 2025 · secondary source

3 · Transparency

limitedWeight 20%

Palantir describes AI risks only in general terms and discloses neither limitations nor error rates for Maven, ImmigrationOS and ELITE.

What is documented
  • The 10-K annual report names risks of AI deployment, the EU AI Act and reputational risks from government contracts1
  • Palantir published responses to the UN Special Rapporteur, Amnesty and the EFF2 3 4
  • According to the 10-K, decisions on the rights of individuals must not be left to computers alone1
The case against
Missing information on high-risk systems
  • There are no system or data cards for Maven, ImmigrationOS and ELITE and no information on performance limits or error rates
  • Accuracy figures for Maven come only from secondary sources8
  • A transparency report and disclosure of high-risk contracts are missing
Disclosure rejected
  • The board rejects an independent report and a human rights impact assessment, citing confidentiality and cost; the proposals failed6 9
Sources for Transparency (9)
  1. Palantir Technologies Inc. / SEC EDGAR: Form 10-K for fiscal year 2025 · 17 February 2026 · primary source
  2. Palantir Technologies: Palantir Response to UN SRHRP Letter – 22 May 2025 · 22 May 2025 · secondary source
  3. Amnesty International: Palantir: Response from Palantir to Amnesty International (AMR 51/0211/2025) · August 2025 · secondary source
  4. Palantir Technologies Inc. (veröffentlicht durch EFF): Reply from Courtney Bowman to the EFF · 6 March 2026 · secondary source
  5. Palantir Technologies (blog): Correcting the Record: Response to the EFF January 15, 2026 Report on Palantir · 3 November 2026 · secondary source
  6. Palantir Technologies Inc. / SEC EDGAR: Form DEF 14A – Proxy Statement 2026 · 24 April 2026 · primary source
  7. Palantir Technologies (product documentation): AIP security and privacy · 3 November 2026 · secondary source
  8. Defense News: Deadly Iran school strike casts shadow over Pentagon's AI targeting push · 24 March 2026 · secondary source
  9. Palantir Technologies Inc. / SEC EDGAR: Form 8-K – voting results of the 2026 annual meeting · June 2026 · primary source

4 · Why the harm rule does not apply

The harm rule does not apply: there is no judgment, regulatory decision or fine against Palantir, and Maven's role in the strike in Minab is known only from media reports. The case still needs review.

5 · Controversies (10)

Further controversies (4)

6 · What remains open

  • Content of the unpublished Pentagon report on the strike in Minab; so far known only through anonymous sources
  • Independent review of AI governance and human rights due diligence
  • Results of customer and use-case reviews for military and immigration deployments
  • Test, validation and error data for high-risk systems
  • Court records in the Medicaid data case and the Zurich Commercial Court judgment in the original

7 · How it was assessed

Cut-off date: 3 November 2026 · 37 searches, 38 fetches.

Primary documents read (10)
  • Form DEF 14A 2026 (S06)
  • Form 8-K voting results AGM 2026 (S07)
  • Form 10-K FY2025, Item 1 and 1A (S10), Item 1C via verbatim reproduction (S11)
  • Palantir Human Rights Policy (S15)
  • Palantir response to the UN Special Rapporteur, 2025-05-22 (S17)
  • Palantir response to the EFF, 2026-03-06 (S14)
  • Amnesty correspondence AMR 51/0211/2025 (S29)
  • Form PX14A6G for AGM 2026 (S01)
  • FedRAMP Marketplace listing (S26)
  • Palantir product documentation on AI ethics, AIP security and data protection (S12, S30, S32)

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