Topics · 03
AI, Personal Data Governance & Cross-Border Compliance
Trace data flows across foundation models, RAG architectures and third-party APIs, auditing legal bases, notice and consent, retention limits, cross-border compliance and exit terms.
How can enterprises protect proprietary data and personal information from unauthorized secondary training and retention?
Key Trigger Scenarios
- Employees entering confidential client data or proprietary documents into public LLMs
- AI systems processing sensitive personal data, biometric traits or user profiling
- Procuring overseas AI models, cloud infrastructure, annotation or compute resources
- Using real production data for model evaluation, fine-tuning or system iteration
Preliminary Due Diligence & Materials
- 01Compile an AI data inventory mapping categories, sensitivity, recipients and cross-border paths
- 02Review vendor terms to prohibit model training, retention and unauthorized onward sharing
- 03Upgrade privacy policies, just-in-time notices, and personal data rights mechanisms
- 04Establish data incident response plans with verified data deletion and exit protocols
Selected Cases
Judicial Trends & Regulatory Standards
Read each case in its procedural context. Related cases may offer comparisons across technologies; they do not establish a single rule for every system.
EU SCHUFA Scoring Case
A credit agency automatically generated credit scores for use by third parties such as banks. The Court of Justice of the European Union held that when a score plays a decisive role in a third party's decision, the scoring activity constitutes automated individual decision-making governed by Article 22 of the GDPR.
Preliminary-ruling judgment interpreting EU law
- Automated scoring falls within GDPR Article 22(1) when a third party relies decisively on the score in making a contractual decision.
- It is prohibited in principle unless an Article 22(2) exception and safeguards apply, alongside other GDPR requirements; an Article 6 legal basis alone is insufficient.
EU Automated Decision-Making Explanation Case
Interpreting GDPR Article 15(1)(h), the CJEU required a concise, transparent and intelligible account of the procedures and principles actually used, enabling the individual to understand and challenge the result. Showing how changes in personal data would change the result is one possible explanatory method.
Preliminary-ruling judgment interpreting EU law
- The account must explain the actual procedures and principles so the person can challenge the decision. A complex formula or exhaustive steps alone may be insufficient.
- Showing the effect of data changes is one possible method, not a mandatory format in every case or a duty to disclose the entire algorithm.
Netherlands SyRI Risk Profiling Case
The Dutch government integrated multi-agency public data to generate fraud risk profiles for social welfare recipients. The court held that the regulatory regime lacked sufficient transparency and auditability, failing to strike a fair balance in its interference with the right to private life.
First-instance civil judgment; government did not appeal
- The safeguards failed to strike a fair balance between fraud prevention and private life under ECHR Article 8(2).
- The court declared SUWI Act Article 65 and Chapter 5a of the implementing decree non-binding, rather than invalidating all government data sharing.
Netherlands Ola Driver Data Access Case
UK drivers sought personal data and explanations concerning Ola’s scoring, fraud detection, income profiles, allocation and deductions. The first-instance court partly allowed access in 2021. In 2023 the Amsterdam Court of Appeal found that the fraud-probability scores, income profiles and allocation at issue had significant effects and ordered further explanations.
First-instance order; partly reversed on appeal on 4 April 2023
- The appellate court partly reversed the first instance, finding that the specified fraud scores, income profiles and allocation were solely automated decisions with significant effects under GDPR Articles 22(1) and 15(1)(h).
- Ola had to explain relevant factors, their aggregate-level weights and information needed to understand the reasoning. Guardian alerts did not meet the same threshold on these facts.
Related Services
Tailored Legal Services for This Scenario
AI & Data Compliance
Map data sources, processing roles, purposes, flows and controls across AI systems.
View Legal Service →03AI Product Compliance
Embed legal review into design, development, procurement, testing, launch, operation and updates.
View Legal Service →05Model & Platform Governance
Review admission, risk tiering, content safety, monitoring, audit and accountability for models, algorithms and platforms.
View Legal Service →06Enterprise AI Governance
Establish enterprise AI planning, tool admission, data permissions, internal policies, training, audit and incident response.
View Legal Service →Regulatory Frameworks
Applicable Regulatory Frameworks
China: PIPL Automated Decision-Making Rules
In force
EU GDPR Automated Decision-Making Rules
In force
Singapore Personal Data and AI Guidelines
Current guidance
Hong Kong AI Data Protection Framework
Current guidance
Research team