AI policy at TGZ Legal

AI Policy at TGZ Legal

Last updated: December 14, 2025
Lawfirm: TGZ Legal - Tegzeș, Burdulea - Societate Civilă de Avocați (SCA)

Why We Use AI

Artificial Intelligence (including generative AI) helps us work more efficiently, structure and analyze information, draft documents, and automate administrative tasks-so our lawyers can focus on legal analysis, strategy, and representation. AI is used under human supervision and within the ethical and legal boundaries of the profession.

Legal Framework and Compliance

  • GDPR / Art. 5(1)(c) - Data Minimization: We process only the personal data strictly necessary for the intended purpose, ensuring security and confidentiality.
  • EU AI Act (Regulation 1689/2024): The European AI regulation entered into force on August 1, 2024, with phased implementation (e.g., AI literacy obligations from February 2025, GPAI model requirements from August 2025, full application from August 2026).
  • Professional Secrecy: Client confidentiality is a fundamental principle; our work complies with law and professional codes.
  • CCBE Guidelines: Confidentiality, competence, independence, and transparency remain central when lawyers use AI.

Tools We Use

We primarily use Microsoft 365 Copilot in its enterprise version (Word, Excel, Outlook, Teams, SharePoint).

  • Microsoft guarantees that prompts, responses, and data accessed via Microsoft Graph are not used to train models and that services comply with GDPR and EU Data Boundary commitments.
  • If we use Azure OpenAI in the future, client data will not be shared with third parties or used for model training without consent.

Note: We do not use public AI services for confidential client matters.

Our AI Usage Principles

Absolute Confidentiality

  • No confidential client data is entered into AI tools without adequate safeguards.
  • All data is processed in accordance with the principles of data minimization, security, accuracy, and storage limitation.

No Model Training on Client Documents

Client documents and information are never used to train or improve AI models.

Respect for Intellectual Property Rights

We carefully check sources and licenses.

Human Oversight & Legal Review

All AI-generated outputs are reviewed and validated by a lawyer before use.

Transparency Toward Clients

Upon request, we offer an approach without AI involvement.

Continuous Risk Assessment

We monitor regulatory developments and update practices accordingly.

Typical Use Cases (with Human Oversight)

  • Drafting and structuring documents: Generation of drafts, summaries, checklists; final version is the lawyer's responsibility.
  • Internal research and summarization: Queries on own documents (within permission limits), extraction of key points.
  • Administrative productivity: Schedules, minutes, working notes, planning.
  • Working translations: Internal indicative translations; official version is verified by lawyer or authorized translator.
  • Large-scale document analysis: Document classification, detection of recurring themes - with legal validation.

Prohibited Uses

  • Entering sensitive data into insecure AI systems.
  • Generating misleading or unlawful content.
  • Automated decisions without human control.
  • Any use that violates professional ethics or confidentiality obligations.

Security and Governance

  • Access control via Microsoft identity and permissions.
  • Encryption at rest and in transit; EU data residency.
  • Logging and internal audits.
  • Ongoing AI literacy training for staff.

Client Rights and Options

No AI Option

Upon request, we will exclude AI from your matter.

Questions or Concerns

Please contact us at office@tgz.legal.

Disclaimer

AI is an internal tool. Professional judgment remains with TGZ Legal. Official documents and advice are always reviewed by a qualified lawyer.

Frequently Asked Questions

Does AI process my personal data?

Only if necessary for service delivery, and always under GDPR principles of data minimization and security.

Are my documents used to train AI models?

No. In Microsoft 365 Copilot, prompts, responses, and Graph data are not used for model training.

Does TGZ Legal comply with the EU AI Act?

Yes. We follow the phased implementation plan (2025-2027) and apply a risk-based approach. Our policy is regularly updated.

What guarantees do I have regarding professional secrecy?

Professional secrecy is legally guaranteed. AI usage does not alter this obligation.

This site uses cookies to improve your experience.
Learn more