Annual Transparency Report

This report explains how Tanqory discloses government/court requests, content takedowns, IP infringement notices, account actions, algorithmic/AI transparency statistics, and data protection/human rights safeguards. It defines scope, metrics, collection methods, reporting cadence, and public/partner communications so they are understandable and auditable.

1) Scope, Purpose, and Definitions

The report covers government/court requests, content/marketplace takedowns, IP infringement notices (copyright/trademark/patent), account suspensions/restrictions, and algorithmic/AI transparency statistics (e.g., number of appeals of automated decisions). It applies to Tanqory’s core platform and products provided to merchants/developers/end users, including ranking/recommendation and decision systems. “Request” means a written request with clear jurisdiction and legal authority. “Account action” means a merchant/developer/user account suspended, restricted, or terminated after review. “Appeal” means a user request to review an automated decision or enforcement action. Standard reporting period is a 12‑month calendar year unless stated otherwise in sub‑reports or real‑time dashboards.

2) Governance and Reporting Cadence

A Transparency Working Group (Legal, Trust & Safety, Security, Data/AI Governance, Privacy/DPO) collects, validates, and approves data before publication. The annual report is published on legal.tanqory.com and announced via email/dashboard (durable medium). Sub‑reports or near‑real‑time dashboards may be shared quarterly with Enterprise partners if contracted. Senior leadership/Board reviews the report at least annually to improve transparency and risk controls. Cadence: main report annually; special updates as needed (e.g., legal changes, new metric categories). If details pose security/legal risk, Tanqory may withhold or aggregate them with reasons stated. If Digital Services Act (DSA) or Platform‑to‑Business (P2B) obligations apply, we will provide the DSA Articles 15–24 disclosures (e.g., reasons for restrictions, complaint outcomes, average monthly active recipients if triggered) and P2B Art.3–4 ranking and restriction reasons to affected merchants, with an internal complaint channel and out‑of‑court options. If Very Large Online Platform/Online Search Engine (VLOP/VLOSE) thresholds are met, we will publish the additional risk assessment/mitigation summaries, recommender transparency, and ad/transparency datasets in the legally required format and cadence.

3) Government & Law Enforcement Requests

We accept written, legally authorized requests and verify warrants/orders. Primary intake: legal@tanqory.com or the official channel stated in the order. Agencies are asked to use secure channels and request only what is legally necessary. We notify affected users/merchants before disclosure where permitted by law and absent a gag order. Reporting breaks down request counts and responses (granted/denied/requested clarification) by country/type, and the number of accounts or data items implicated. We do not provide backdoor access or break end‑to‑end encryption except where explicitly required by law. MLAT processes are followed where applicable. We disclose incomplete/denied requests with policy/legal reasons (e.g., lack of authority, unclear jurisdiction, overbroad scope).

4) Content and Account Actions

We report notices and actions by category (illegal/policy‑violating content, security/phishing/malware, prohibited/high‑risk goods, fraud/scams, local‑law violations) and account suspensions/restrictions. Human‑in‑the‑loop is used for high‑impact cases, with appeal channels. Appeals and outcomes (upheld/reversed/mitigated) are reported separately. We include response/closure time metrics (e.g., average intake‑to‑close, SLA adherence) to show enforcement efficiency. We clarify platform vs. merchant roles when content/items originate from third‑party stores. Under DSA/P2B, we will provide reason statements for demotion/removal to affected merchants/users, including policy violated, main automated signals considered, and how to remedy; we will disclose complaint handling timelines and outcomes (dismissed/upheld/modified) and any out‑of‑court dispute resolution used.

5) IP Infringement Requests

We intake through published channels (DMCA/Trademark/Patent forms) and validate completeness. Reporting shows counts and outcomes by IP type, and the number of URLs/items/accounts affected. Counter‑notices/appeals are reported separately, with average response/closure times. Trends (e.g., most‑reported content types, reversal rates) are summarized while respecting privacy and NDAs.

6) Algorithmic/AI Transparency

We report counts of notices/appeals related to automated decisions, ranking, or recommendations; the number of published Model/Service Cards; requests for explanations/appeals and their outcomes (upheld/modified/reversed). We list major model/provider updates announced to users, and summarize legally reportable AI incidents (if any) per SLA/regulatory rules. We note watermarking/provenance updates for AI‑generated content. Reporting aligns with the AI Ethics Framework and Algorithmic Transparency Policy. For DSA/P2B compliance, we also disclose main ranking parameters and relative importance for recommender systems in scope, any significant changes to those parameters, and counts of merchant complaints about ranking/removal with resolutions. If VLOP/VLOSE, we will include systemic risk topics, mitigations, and independent audit summaries as required.

7) Data Collection and Accuracy

Data is drawn from governed ticketing/incident/appeal systems with audit trails and validated by Legal/Trust & Safety before release. We count by “event” and “account/item” to avoid confusion; if sampling/aggregation is used, methods are described. Any redactions are for security/legal constraints and are explained. Metrics are defined in an appendix for year‑over‑year comparability. Standard metric set (minimum):

• Government/court requests: count by country/type; granted/denied/clarified; accounts/data items affected; notice provided (yes/no).
• Content/account actions: notices by category; actions taken; appeals received/resolved; upheld/reversed; median response/closure times.
• IP: notices by IP type; URLs/items/accounts affected; counter‑notices; reversals; median response/closure times.
• Algorithmic/AI: explanation/appeal requests; outcomes; Model/Service Cards published; major model/provider changes; AI incidents reportable.
• DSA/P2B (where applicable): merchant complaints; outcomes; reason statements; ranking parameter changes; monthly active recipients (if required).
  Formats: human‑readable PDF/HTML plus CSV/JSON for key tables where feasible to support external review.

8) Privacy and Data Protection

No personal data is published unless legally required; data is redacted/anonymized. We comply with Privacy Policy, DPA, and data laws (GDPR/UK GDPR, CPRA, PDPA SG, PIPEDA, AU Privacy Act) in publishing statistics. Where law allows, we notify users about government requests unless it risks safety or violates orders. Access to report data is role‑based with access logs; supporting records follow the Data Retention Policy. Retention: keep source records (notices, reason statements, appeals, outcomes, timestamps, consent/notice logs) for at least the statutory limitation or contractual period applicable in each region; minimum 2 years unless law requires longer (e.g., DSA evidence for disputes). Redactions and aggregations are documented to preserve auditability.

9) Regional Requirements

EU/UK: comply with ePrivacy/PECR and GDPR/UK GDPR for notices and rights to appeal automated decisions; prepare for AI Act transparency/serious incident reporting. US: comply with CPRA/US state privacy for “Do Not Sell/Share” disclosure and GPC signals where applicable, and FTC fair‑communication expectations. Canada: PIPEDA/provincial transparency requirements. Singapore: PDPA and consumer/ad rules. Australia: Privacy Act and ACL. Other markets: add local‑law text as regulations take effect. Timelines for notice/publication are adjusted to applicable law. Under DSA/P2B, we will include ranking explanations, merchant notice for demotion/removal, internal complaint handling, and out‑of‑court options if thresholds/roles apply; if VLOP/VLOSE thresholds are met, additional transparency reporting will follow the DSA. Children/young users: apply age‑appropriate design/online safety where applicable (e.g., UK AADC, EU/US child codes); avoid profiling/ads or non‑essential cookies without valid consent; use simple notices. Local language requirements (e.g., Quebec French) are honored for affected audiences.

10) Publication, Updates, and Contact

Annual report is published on legal.tanqory.com; key metrics may appear on dashboards. Users/merchants/developers are notified via email/dashboard when reports or major updates are available. Clarifying edits update version/date without reducing user rights; material changes are announced with reasonable notice. Contact for inquiries or additional information: legal@tanqory.com or contract‑specified channels. Enterprise customers may request bespoke reports if agreed in SLA/Order Form. English controls if translations differ. We target publication in open, accessible formats (HTML/PDF plus CSV/JSON tables) and aim to announce publication dates in advance. If systemic risk reports (DSA VLOP/VLOSE) or recommender transparency files are required, they will be linked or hosted alongside the annual report.

EU DSA Article 24 metrics framework

The Annual Transparency Report intends to cover the following categories of metrics on a going-forward basis, mirroring the disclosure framework of Article 24 of Regulation (EU) 2022/2065 (Digital Services Act):

• Number of orders received from Member States' authorities under Article 9, broken down by Member State and type of illegal content.
• Number of notices submitted under Article 16, including statistics on action taken on the basis of those notices, broken down by category and accuracy of decisions.
• Number of complaints received through the internal complaint-handling system under Article 20, with statistics on decisions, average time taken, and reversals.
• Out-of-court dispute settlements under Article 21 — number of disputes, outcomes, and median time to complete the settlement.
• Suspensions of recipients of the service under Article 23 — number and grounds, including suspensions for manifestly illegal content, manifestly unfounded notices, or manifestly unfounded complaints.
• Use of automated means in content moderation — categories of automated tools used, the qualitative description of their purposes, accuracy and error-rate estimates where measurable, and safeguards applied.

The first reporting period will be the first full calendar year following the date the applicable DSA reporting obligation takes effect.

---

Document Details