top of page

Politiques et contrats juridiques

Traduction fournie à titre informatif. En cas de divergence, la version anglaise fait foi.

AI Act Compliance Policy

Last updated: 7 November 2026 (version 2.0)
Company: InteliATE Ltd (Company No. 15514345), registered in England and Wales
Registered office: 35 Cedars Close, Hendon, London NW4 1TR, United Kingdom
Policy owner: David Pinto, Founder and CEO (Compliance Lead)
Contact: contact@inteliate.com (subject: AI Act)

Avis : cette politique est publiée en anglais, langue de travail d’InteliATE Ltd. La version anglaise fait foi. Pour obtenir une traduction, écrivez à contact@inteliate.com.

1) Purpose and scope

This policy sets out how InteliATE meets Regulation (EU) 2024/1689 (the EU AI Act), as amended by Regulation (EU) 2026/1744 (the Digital Omnibus on AI), for the AI systems it builds and supplies. It applies to the directors of InteliATE Ltd, to every independent professional who works for it through their own company, and to development partners and subcontractors who work on those systems.

Status today: InteliATE builds to the high-risk requirements ahead of the dates on which they apply. It has not yet completed a conformity assessment, applied CE marking or registered a system in the EU database for any product, because those steps are not yet due for stand-alone Annex III systems (2 December 2027). Nothing in this policy states that a product is certified.

1.1 Our role

InteliATE is the provider of its AI systems and their original equipment manufacturer (OEM). Our customers are deployers. A customer that changes a system's intended purpose, or makes a substantial modification, can become a provider. Our contracts say so.

1.2 Classification

The Compliance Lead keeps a written classification record for each product and each customer use. It records the intended purpose, any Annex III category, whether an Article 5 prohibition applies, and the countries where the system is used. The record is reviewed before each new customer, new use and major release. It is available to customers and regulators on request.

We currently treat these systems as high-risk:

  • Fusion AI, including the National Data Fusion Centre and Tactical Investigation Manager, when used by police in investigations (Annex III, point 6).

  • Triage AI, when used by police to evaluate forensic evidence (Annex III, point 6).

  • OSINT Hub, when a customer uses it to profile persons (Annex III, point 6).

  • Passenger Risk Screening, for border and passenger checks (Annex III, point 7), and under point 1 where facial recognition is used.

  • Personnel Intelligence, when a customer uses it for employment decisions about workers or candidates (Annex III, point 4).

  • Cyber modules (Network Security, Data Protection, Exposure Management and Security Operations), when a customer uses them to assess workers or other persons.

  • Container Intelligence and X-ray Intelligence, when a customer uses them to assess persons.

We do not supply a use until it has been classified.

1.3 Obligations for high-risk systems

For stand-alone Annex III systems, the high-risk obligations apply from 2 December 2027. For systems embedded in regulated products (Annex I), they apply from 2 August 2028. Until then we build to the same requirements, which are:

  • Risk management across the product life cycle (Article 9).

  • Data governance, including the origin, suitability and bias testing of data (Article 10).

  • Technical documentation (Article 11 and Annex IV) and automatic logging (Article 12).

  • Instructions for use that let deployers read and use outputs correctly (Article 13).

  • Human oversight designed into the product (Article 14).

  • Accuracy, robustness and cybersecurity, with stated metrics (Article 15).

  • A quality management system (Article 17).

  • Conformity assessment, EU declaration of conformity and CE marking where required (Articles 43, 47 and 48).

  • Registration before a system is placed on the EU market (Article 49). Law enforcement and migration systems are registered in the non-public part of the EU database.

  • Post-market monitoring (Article 72) and reporting of serious incidents (Article 73).

  • An authorised representative established in the EU, appointed before we place any system on the EU market from the UK (Article 22).

1.4 Prohibited uses

We do not build, supply or support any system for a practice that Article 5 prohibits. That includes:

  • social scoring of people (Article 5(1)(c));

  • predicting criminal offences from profiling alone (Article 5(1)(d));

  • building or expanding facial recognition databases through untargeted scraping of images from the internet or CCTV (Article 5(1)(e));

  • emotion recognition in workplaces and schools (Article 5(1)(f));

  • biometric categorisation that infers race, political opinions, religion, sexual orientation and similar traits (Article 5(1)(g));

  • real-time remote biometric identification in publicly accessible spaces for law enforcement, except the narrow exceptions in Article 5(1)(h). We will not support any use under those exceptions without written advice from counsel;

  • generating non-consensual sexually explicit images or child sexual abuse material, which Regulation (EU) 2026/1744 adds to Article 5 from 2 December 2026.

1.5 Duties already in force

  • AI literacy (Article 4), which has applied since 2 February 2025. Our people who build or use AI systems must understand their limits, and we train customers on the limits of each system and on human oversight at deployment.

  • General-purpose AI model obligations (Articles 51 to 56), which have applied since 2 August 2025. InteliATE does not itself provide a general-purpose AI model. If it ever does, it will meet those articles.

  • Transparency duties (Article 50), where a system interacts with people or generates synthetic content. We design our products to support those duties and tell customers what they must do.

  • In the UK, there is no AI-specific statute. UK GDPR rules on automated decision-making, as amended by the Data (Use and Access) Act 2025, apply to customers that use our products to take decisions about people.

1.6 Case isolation and model updates

  • Each investigation runs in its own isolated workspace. Its AI state and outputs are scoped to that case and are not shared with other cases.

  • Analyst feedback applies only within the case where it was given.

  • Customer data is not used to train a model that is shared between customers.

1.7 Human oversight

AI output is advisory. A person makes every decision about another person. Our products are designed so that a person reviews an output before any decision about a person is taken. Documentation states the limits and known error modes.

1.8 Serious incidents

Anyone who suspects a serious incident reports it to the Compliance Lead without delay. The Compliance Lead decides on reporting to the market surveillance authority within the deadlines in Article 73, which are shorter for the most serious cases, and informs the customer.

1.9 Roles

  • Executive owner and Compliance Lead: David Pinto, Founder and CEO.

1.10 Records and review

Technical documentation is kept for 10 years after a system is placed on the market, as Article 18 requires. Logs, the classification record and AI literacy records are kept as well. We review this policy whenever a product, a customer country or the law changes, and at least every 12 months.

Sources: Regulation (EU) 2024/1689 (https://eur-lex.europa.eu/eli/reg/2024/1689/oj), as amended by Regulation (EU) 2026/1744 (Digital Omnibus on AI, Official Journal of the European Union, 24 July 2026).

bottom of page