EU AI Act compliance

Last updated: 16 June 2026

At dzply, we provide the foundational algorithmic infrastructure, private GPU hosting, and workflow automation engines that power the creator economy and modern digital workforces. We are committed to strict legal alignment with the European Union Artificial Intelligence Act (EU AI Act) and the General Data Protection Regulation (GDPR).

1. Architectural Sovereignty & GDPR Compliance

  • Private GPU Infrastructure: All proprietary Large Language Model processing—including our deployed DeepSeek model architectures—runs entirely on dzply's privately owned or leased dedicated GPU infrastructure.
  • Zero-Leakage Guarantee: Your inputs, contextual prompts, and generated outputs stay completely within our secure environment. Your data is never exposed to external third-party AI cloud vendors or used to train public base models.
  • Third-Party Routing: When users explicitly opt to use non-dzply third-party APIs, data transfers are strictly isolated under Data Processing Addendums (DPAs) that prohibit model training on user data.

2. High-Risk AI Classification & Joint Responsibility (Recruitment Tools)

Under Annex III of the EU AI Act, AI systems used for recruitment, candidate screening, work task allocation, and performance evaluation are legally classified as High-Risk AI Systems.
  • Shared Liability: Because dzply provides the infrastructure allowing users to deploy autonomous recruitment agents, we provide built-in tools to help users satisfy their legal obligations under Article 26 of the AI Act.
  • Mandatory Risk Management: Any user building recruitment or employment-related agents on dzply must complete our internal High-Risk Configuration Protocol, enabling strict quality control, bias monitoring, and human-in-the-loop validation.
  • Fundamental Rights Impact Assessments (FRIA): Our infrastructure is technically optimized to generate the necessary logs and data summaries required for users to submit mandatory FRIAs to EU authorities before deployment.

3. Generative Media and Agent Transparency

  • Synthetic Media Labeling: In accordance with Article 52 of the AI Act, all audio, visual, or textual media generated via dzply's infrastructure contains hidden, machine-readable digital watermarks identifying it as AI-generated.
  • Interaction Disclosures: Our workflow automation and AI workforce systems are hardcoded to display clear disclosures to external end-users, ensuring they are always aware they are interacting with an AI agent.

4. Technical Safeguards & Data Quality

  • Bias Mitigation: We actively audit our hosted DeepSeek model implementations to reduce systemic algorithmic bias, particularly concerning protected characteristics under EU law (gender, race, age, disability).
  • Continuous Logging: dzply automatically maintains unalterable, chronological system logs of high-risk workflows to guarantee system predictability, traceability, and post-market monitoring capabilities.
  • Human-in-the-Loop Override: All automation sequences feature mandatory "kill-switch" mechanisms and manual override parameters, ensuring absolute human control over autonomous systems.

5. Prohibited Practices

dzply strictly blocks, monitors against, and bans the creation of AI tools intended for prohibited EU practices. This includes untargeted scraping of facial images, workplace emotion recognition, and subconscious behavioral manipulation.

6. Compliance Contact & Documentation Access

To request technical documentation, data subject access requests, or compliance logs for an AI Act regulatory audit, contact our legal office at: hello@dzply.com.