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

Expedia Group

USAAdoptersAssignment FullData quality Medium

Key messagesolid AI Footprint

Expedia uses artificial intelligence in customer service, search and pricing and against fraud, and describes an internal council that is meant to review risky applications before they are introduced. Users are told that answers may contain errors, but the third-party models used are not named, and no independent testing of the systems was found.

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

Opportunities

  • Safetysolid

    A cyber and privacy programme based on NIST CSF and PCI DSS is documented, with an incident response plan tested annually, third-party requirements, a 98 % training rate and board oversight (S04, S05, S03), as well as certifications under Global CBPR and the Data Privacy Framework (S10).

    A cyber and privacy programme based on NIST CSF and PCI DSS is documented, with an incident response plan tested annually, third-party requirements, a 98 % training rate and board oversight (S04, S05, S03), as well as certifications under Global CBPR and the Data Privacy Framework (S10). On abuse protection, the DSA report documents automated detection of inauthentic content and suspicious accounts with human review (S12). On the reliability of the generative systems there are only management statements on evaluations and red-teaming (S14, S15); an independent test found the Hotels.com chatbot error-prone (S16), Romie was discontinued (S15), and user reports describe refund requests that were never created (S18). This counter-evidence is minor but prevents a higher rating.

  • Transparencysolid

    In its privacy statement Expedia discloses the areas where it uses AI, including pricing, ranking, recommendations, chatbots and review summaries, together with rights in relation to automated decisions (S09, S10).

    In its privacy statement Expedia discloses the areas where it uses AI, including pricing, ranking, recommendations, chatbots and review summaries, together with rights in relation to automated decisions (S09, S10). The AI terms label answers as machine-generated and point out possible errors (S08); the DSA report describes the automated moderation but gives no error rates (S12). Information on the third-party models used is limited (only “third-party large language models”, S08), as is information on how the pricing algorithms work, for which only a statement to the media is available (S33).

Risks

  • Responsibilitylimited

    The 2025 Global Impact Report describes an AI policy that requires formal impact assessments and documented model performance for high-risk applications, and a Responsible AI Council in place since 2023 made up of technology, legal, privacy and data science, which reviews AI use cases (S04); this …

    The 2025 Global Impact Report describes an AI policy that requires formal impact assessments and documented model performance for high-risk applications, and a Responsible AI Council in place since 2023 made up of technology, legal, privacy and data science, which reviews AI use cases (S04); this documents a review procedure published by the company. However, AI-specific oversight by the board, a committee or executive management is not demonstrated with grade A or B evidence: the Audit Committee oversees product security and cyber risks without mentioning AI (S05), the council's composition at executive level is not documented, and the Chief AI and Data Officer is documented only through press reports and conference statements (S13, S14). Without AI-specific oversight the pillar remains “limited”; in addition, the full text of the policy, AI metrics, external audits and certifications are missing, while complaint channels are documented concretely in the DSA report (S12).

  • Risk area

    AI customer service affecting refunds and rebookings

  • Risk area

    Algorithmic and personalised pricing based on personal data

  • Risk area

    Fraud detection, account blocks and automated content moderation

  • No serious controversy recorded.

What lies behind the assessment

Research of 3 November 2026
45searches
37pages and documents read
9primary documents
3controversies checked
6open points, named openly
The full research: pillars, evidence, controversies and open points
AI Footprint · Adopter · As of 3 November 2026
Expedia GroupOverall rating: solid

Expedia has risky AI applications reviewed by an internal council and discloses where it uses AI in its privacy statement. The policy itself, the third-party models used and independent testing of the AI customer service are not published.

Responsibility

40 %
limitedsolidstrong

An AI policy with impact assessments is described, but AI-specific oversight is missing.

  • +Responsible AI Council reviews high-risk applications before scaling Source ↗
  • +Complaint channels documented with figures in the DSA report Source ↗
  • –AI policy available only as a summary in the Impact Report Source ↗
  • –No AI-specific oversight by the board or executive management documented Source ↗
  • –No AI metrics and no external audit of AI systems
All sources for this pillar (13)

Safety

40 %
limitedsolidstrong

A cyber programme based on NIST is documented, but the reliability of the AI customer service is not.

  • +Cyber programme based on NIST CSF and PCI DSS with a tested response plan Source ↗
  • +Automated detection of inauthentic content with human review Source ↗
  • –Independent test found the Hotels.com chatbot error-prone Source ↗
  • –Users report refund requests never created by the AI support Source ↗
  • –No error rates or red-teaming results published
All sources for this pillar (14)

Transparency

20 %
limitedsolidstrong

Where AI is used is disclosed, but the third-party models used are not named.

  • +Privacy statement lists AI uses from pricing to chatbots Source ↗
  • +AI answers labelled as machine-generated and possibly inaccurate Source ↗
  • –Language models named only as “third-party large language models” Source ↗
  • –On AI pricing only a statement to the media Source ↗
  • –No error rates for the automated moderation Source ↗
All sources for this pillar (9)
What is missing for “strong”

These steps are in the company’s own hands.

1Responsibility
Anchor AI oversight and disclose the policy

Make a committee or executive management explicitly responsible for AI, publish the AI policy and provide figures on the use cases reviewed.

2Safety
Have the AI customer service tested independently

Publish independent tests, red-teaming results and error and escalation rates of the Virtual Agent.

3Transparency
Explain the models and pricing logic

Name the foundation models used and explain how the pricing algorithms work and what data they rely on.

Context: 101 companies assessed

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

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
Faulty answers from the AI customer service

User reports say the AI customer service confirmed refund requests that were never created; no independent testing or error rates are published.

How thoroughly assessed
36 Evidence

9 primary documents read, 45 searches.

Read the full reasoning and all sources
Expedia Group · Full reasoning

Internal AI council documented, but no independent testing of the AI customer service

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

1 · Responsibility

limitedWeight 40%

Expedia describes an AI policy with impact assessments and an internal AI council, but neither the board nor executive management is demonstrably responsible for AI explicitly.

What is documented
  • Since 2023 a Responsible AI Council of technology, legal, privacy and data science has reviewed certain AI use cases3
  • The AI policy is based on the OECD AI Principles and requires formal impact assessments for high-risk applications3
  • The first Chief AI and Data Officer was appointed in December 2025; he describes risk-based review gates with evaluations and red-teaming4 7
  • The Audit Committee oversees product security and cyber risks; the CISO and CTO report at least once a year1 2
  • According to the DSA report there were 781 complaints and 477 reversals, with a median handling time of one day11
  • The privacy statement grants a right to a manual decision and to challenge automated decisions12 13
The case against
Self-description without figures
  • The AI policy is not published; the information comes from the Impact Report and from management appearances3 7 5
  • The number of use cases reviewed by the council and other AI metrics are not available
  • No external audit or certification of AI systems was found
Gaps in responsibility and control
  • No board committee and no documented executive management body is explicitly responsible for AI1
  • No approval procedure for purchased language models is published; the AI terms name no channel for reporting AI errors10
Sources for Responsibility (13)
  1. Expedia Group, Inc. / SEC: Definitive Proxy Statement (DEF 14A) 2026 · 29 April 2026 · primary source
  2. Board Cybersecurity (board-cybersecurity.com): Expedia Group, Inc. 10-K Item 1C. Cybersecurity – 2026-02-12 · 12 February 2026 · secondary source
  3. Expedia Group, Inc.: Expedia Group 2025 Global Impact Report · 31 December 2025 · primary source
  4. Short Term Rentalz: Expedia Group appoints first chief AI and data officer · 2 December 2025 · secondary source
  5. Expedia Group (Partner-Blog): AI at Explore 2026: what partners need to know · 8 June 2026 · secondary source
  6. Hospitality.today: Expedia takes a measured path toward agentic AI · 2 April 2025 · secondary source
  7. VentureBeat: Evals are the new PRD, Expedia AI chief tells VB Transform 2026 · 21 July 2026 · secondary source
  8. Expedia Group, Inc. / SEC: Form 10-K for fiscal year 2025 · 12 February 2026 · primary source
  9. Expedia Group, Inc. / SEC: Form ARS 2025 (annual report to shareholders) · 3 November 2026 · primary source
  10. Expedia, Inc.: AI Terms (terms of use for AI tools) · 3 November 2026 · primary source
  11. Expedia, Inc.: DSA Transparency Report Expedia, Inc. (17.02.2024–31.12.2024) · 31 December 2024 · primary source
  12. Expedia, Inc.: Privacy Statement (expedia.ca) · 3 November 2026 · primary source
  13. Expedia Group: Expedia Group Privacy Statement (legal.expediagroup.com) · 30 January 2026 · primary source

2 · Safety

solidWeight 40%

Expedia documents a security and privacy programme with certifications, but on the reliability of its generative systems there are only management statements, and an independent test and user reports point to errors.

What is documented
  • The programme follows NIST CSF, PCI DSS and NIST Privacy by Design; the incident response plan is tested annually4 8
  • Expedia is certified under Global CBPR and the Data Privacy Framework and undergoes PCI DSS assessments10
  • According to the Impact Report there were no material data breaches in 2025; the training rate is 98 %4
  • The DSA report shows 7,867 automatically detected actions; all notices are reviewed by people11
  • After unsupported chatbot answers, Expedia rebuilt its evaluation framework and data basis and discontinued the Romie assistant2
  • AI agents may recommend offers but may not book on their own1
The case against
Errors in the AI customer service
  • A hands-on test by Skift found the Hotels.com chatbot error-prone, with inaccurate recommendations and false information5 6
  • App reviews report that the AI customer service confirmed refund requests that were never created; scale and cause are not documented7
No independent testing
  • Independent tests, red-teaming results or error rates of the Virtual Agent are not published
  • Item 1C on cybersecurity does not mention AI and is available only in a secondary account8
  • The 10-K itself warns of fraud through agentic bookings with weak consent controls14
Sources for Safety (14)
  1. VentureBeat: Evals are the new PRD, Expedia AI chief tells VB Transform 2026 · 21 July 2026 · secondary source
  2. Hospitality.today: Expedia learned where AI breaks. It wasn't the model · 16 June 2026 · secondary source
  3. Travel Weekly (Australia): Expedia CEO warns that end-to-end AI chatbots don't work · 20 May 2026 · secondary source
  4. Expedia Group, Inc.: Expedia Group 2025 Global Impact Report · 31 December 2025 · primary source
  5. Skift: Hotels.com Has a New Chatbot Trip Planner. Here's How it Works · 20 May 2025 · primary source
  6. Horwath HTL: Case Study: Hotels.com AI Chatbot Trip Planner · 20 May 2025 · secondary source
  7. MWM: Expedia v2026.18: AI chatbot fails to initiate claims · 10 May 2026 · secondary source
  8. Board Cybersecurity (board-cybersecurity.com): Expedia Group, Inc. 10-K Item 1C. Cybersecurity – 2026-02-12 · 12 February 2026 · secondary source
  9. Expedia Group, Inc. / SEC: Definitive Proxy Statement (DEF 14A) 2026 · 29 April 2026 · primary source
  10. Expedia Group: Expedia Group Privacy Statement (legal.expediagroup.com) · 30 January 2026 · primary source
  11. Expedia, Inc.: DSA Transparency Report Expedia, Inc. (17.02.2024–31.12.2024) · 31 December 2024 · primary source
  12. Expedia, Inc.: Privacy Statement (expedia.ca) · 3 November 2026 · primary source
  13. Expedia, Inc.: AI Terms (terms of use for AI tools) · 3 November 2026 · primary source
  14. Expedia Group, Inc. / SEC: Form ARS 2025 (annual report to shareholders) · 3 November 2026 · primary source

3 · Transparency

solidWeight 20%

Expedia lists the areas where it uses AI in its privacy statement and Seals AI answers as machine-generated, but names neither the language models used nor how its pricing algorithms work.

What is documented
  • The privacy statement names AI use in pricing, ranking, recommendations, chatbots and review summaries1 5
  • The AI terms label answers as machine-generated and point out inaccurate answers and unverified sources2
  • Expedia rules out automated decisions with significant effects and grants a right to a manual decision1 5
  • The DSA report describes the automated moderation with human review3
The case against
Missing model and system information
  • The foundation models used are not named; no system description of the Virtual Agent was found2 7
  • Whether chat data is used to train models is not explicitly addressed2 1
  • Skift's test does not describe any labelling in the Hotels.com chatbot4
Pricing and error rates
  • On AI-supported pricing there is only the statement to the media that prices are not raised on the basis of user data6
  • The US House Oversight Committee requested documents on Expedia's pricing policies; no allegations were made9 6
  • The DSA report gives no error rates for the moderation models3
Sources for Transparency (9)
  1. Expedia, Inc.: Privacy Statement (expedia.ca) · 3 November 2026 · primary source
  2. Expedia, Inc.: AI Terms (terms of use for AI tools) · 3 November 2026 · primary source
  3. Expedia, Inc.: DSA Transparency Report Expedia, Inc. (17.02.2024–31.12.2024) · 31 December 2024 · primary source
  4. Skift: Hotels.com Has a New Chatbot Trip Planner. Here's How it Works · 20 May 2025 · primary source
  5. Expedia Group: Expedia Group Privacy Statement (legal.expediagroup.com) · 30 January 2026 · primary source
  6. PhocusWire: US Oversight Committee questions online travel companies on AI pricing · March 2026 · secondary source
  7. Horwath HTL: Case Study: Hotels.com AI Chatbot Trip Planner · 20 May 2025 · secondary source
  8. Expedia Group, Inc.: Expedia Group 2025 Global Impact Report · 31 December 2025 · primary source
  9. U.S. House Committee on Oversight and Government Reform: Comer Investigates Use of Artificial Intelligence to Set Prices for Consumers · 5 March 2026 · primary source

4 · Why the harm rule does not apply

The harm rule does not apply: within the 36-month window there is no judgment, no authority decision, no fine and no pattern of serious incidents involving its own AI. The chatbot cases are minor, and the parliamentary inquiry contains no finding.

5 · Controversies (3)

6 · What remains open

  • Full text of the Responsible AI policy and the number of use cases reviewed by the council
  • AI-specific responsibility at board, committee or executive management level
  • Independent testing, red-teaming results or error rates of the AI customer service
  • Names of the foundation models used and approval procedures for purchased AI
  • How the algorithmic pricing works and what data it relies on
  • Item 3 Legal Proceedings of the 2025 10-K and Part II Item 1 of the 2026 10-Q in full

7 · How it was assessed

Cut-off date: 3 November 2026 · 45 searches, 37 fetches.

Primary documents read (9)
  • Form 10-K 2025 (Items 1 and 1A, in part)
  • Form ARS 2025 (Item 1A, in part)
  • Proxy Statement DEF 14A 2026 (up to shareholding disclosures)
  • Form 10-Q Q2 2026 (in part)
  • Global Impact Report 2025 (complete, incl. pp. 30–31 Privacy, Cybersecurity and AI)
  • DSA Transparency Report Expedia, Inc. 2024
  • AI Terms, Privacy Statement (expedia.ca), Expedia Group Privacy Statement (as of 2026-01-30)
  • Press release and letter from the House Oversight Committee (2026-03-05, 2026-09-09)
  • European Commission communication on commitments by online travel agencies (2025-07-31)

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