Information provided pursuant to Law 23 September 2025, No. 132 and EU Regulation 2016/679 (GDPR)
Array adopts a cutting‑edge approach to integrating Artificial Intelligence (“AI”) systems, using them exclusively as tools for support, optimisation, and document analysis. This notice describes the technical protocols adopted to guarantee the highest confidentiality and the integrity of the professional mandate. The decision to use—or not use—these tools and their outputs rests solely on the personal judgement of each professional. No externally‑relevant activity is performed automatically.
1. Local Architecture and Confidentiality (On‑Premise)
Unlike typical cloud solutions, Array has implemented a private computing infrastructure that safeguards data sovereignty:
- Local Models (LLM) – Analyses run on proprietary physical machines, using models that do not transmit data to remote servers nor rely on external API keys. Policy sull’Utilizzo di Sistemi di Intel…
- RAG & Vector Databases – We employ Retrieval‑Augmented Generation (RAG) systems linked to private vector databases. This enables secure querying of archived documents without ever leaving the security perimeter of the Array professionals’ studio. Policy sull’Utilizzo di Sistemi di Intel…
- MCP (Model Context Protocol) Systems – Document‑collection queries are performed via MCP protocols that isolate the informational context, ensuring the AI operates only on data authorised for the specific task.
2. Selected Third‑Party Tools
For specific linguistic or jurisprudential‑research needs, Array uses only professional platforms with enterprise‑grade Service Level Agreements (SLAs):
- DeepL Pro – Used for translation and refinement of text segments; the existing contract explicitly excludes the use of submitted texts for model training. Policy sull’Utilizzo di Sistemi di Intel…
- Certified Legal Databases – AI‑enhanced legal databases (e.g., Giuffrè Francis Lefebvre) are employed for statistical analysis of case law.
3. Human Supervision and Professional Ethics
AI usage is strictly limited to instrumental, organisational, or preliminary‑analysis activities.
- Continuous Verification – Every AI‑generated output undergoes a critical, thorough, and complete review by Array’s professionals, both for source validation and legal merit (Human‑In‑The‑Loop).
- Intellectual Value – Professional responsibility and legal strategy remain exclusively human. In compliance with deontological duties, billed work reflects the original intellectual contribution and the time spent supervising and transforming data into legal advice.
4. Transparency Statement
By engaging the firm, the client confirms that they have been informed about the methods of AI usage, which are intended solely to enhance the quality and timeliness of the professional service while fully respecting privacy and confidentiality. Policy sull’Utilizzo di Sistemi di Intel…
This document was drafted and approved by Array’s professionals on 26 March 2026.