Compliance Framework under Article 4 of Regulation (EU) 2024/1689 – EU Artificial Intelligence Act
Effective Date: 30 September 2026
1. Purpose
SUGUD – Advocacy for Families & Human Rights recognizes Artificial Intelligence (AI) as a technological tool that can support research, documentation, translation, communication and administrative work when used responsibly and under meaningful human oversight.
This policy establishes SUGUD’s organizational framework for AI literacy, responsible AI use, human oversight and verification, with particular regard to Article 4 of Regulation (EU) 2024/1689 (EU Artificial Intelligence Act), as amended.
SUGUD adopts this policy as part of its organizational measures to support appropriate AI literacy among persons using AI systems on behalf of the organization.
2. Academic and Practical AI Literacy
The founder of SUGUD, Ahmed Homos, holds a Bachelor’s degree in Information Systems.
The qualification has been formally assessed by the German Central Office for Foreign Education (Zentralstelle für ausländisches Bildungswesen – ZAB) and recognized as corresponding to a German higher-education qualification at Bachelor level.
This academic background, together with practical experience in digital information systems and AI-assisted tools, forms part of SUGUD’s organizational approach to AI literacy, human oversight and responsible use of artificial intelligence.
Academic qualifications alone are not treated as proof of compliance. SUGUD therefore combines relevant knowledge and experience with practical safeguards, human verification and continuing awareness of the capabilities and limitations of AI systems.
3. AI Literacy Principles
SUGUD considers responsible AI literacy to include an appropriate understanding of:
- the capabilities and limitations of AI systems;
- the possibility of inaccurate, incomplete, outdated or fabricated AI outputs;
- the importance of verification against primary sources;
- human oversight and accountability;
- privacy, confidentiality and data-protection considerations;
- the distinction between AI-assisted processing and documentary evidence; and
- the particular risks associated with the use of AI in legal, judicial and human-rights documentation.
The level and nature of human oversight applied should reflect the context, purpose and potential consequences of the relevant AI-assisted activity.
4. Permitted Uses of AI
SUGUD may use AI-assisted software and systems for supporting functions including:
translation and linguistic assistance; document formatting and structuring; summarisation; organization and classification of information; comparison of documents; research assistance; administrative and office functions; identification of potentially relevant legal provisions; and preparation or linguistic refinement of draft materials.
AI remains an assisting tool in these activities and does not replace appropriate human review.
5. Special Safeguards for Legal and Human-Rights Reporting
Because SUGUD works with judicial, administrative and human-rights documentation, enhanced verification principles apply to formal reports and submissions.
AI output is not evidence.
Where SUGUD reports an alleged judicial, administrative, institutional or human-rights violation, the evidentiary basis must remain the underlying source material, including, where applicable:
court decisions, official correspondence, procedural records, administrative documents, expert reports, audiovisual material and other identifiable documentary evidence.
AI may assist in translating, organizing, comparing, summarizing or analysing such material.
It does not replace the underlying evidence.
Material names, dates, file numbers, quotations, legal provisions and factual assertions should therefore be checked against the relevant source documentation before being relied upon in formal reporting.
6. Human Oversight and Legal Assessment
SUGUD does not delegate final responsibility for its reports, publications or formal submissions to an AI system.
AI systems do not independently determine whether a person, court, public authority or institution has committed a legal or human-rights violation.
Where such a conclusion is published by SUGUD, it must result from human assessment of the available documentary evidence and relevant legal framework.
AI-generated analysis or suggestions may therefore be accepted, rejected, corrected or independently verified before use.
7. Accuracy and Hallucination Safeguards
SUGUD recognizes that generative AI systems may produce inaccurate, incomplete, outdated or fabricated information, commonly referred to as AI hallucinations.
Consequently, AI-generated information concerning material facts, legal authorities, judicial decisions, quotations, dates, names or documentary evidence should not be relied upon solely because it was produced by an AI system.
Where primary documentation is available, the primary documentation prevails.
Where an inconsistency cannot be resolved from the available evidence, it should be identified as an unresolved issue rather than converted into an unsupported factual conclusion.
8. Data Protection and Confidentiality
Particular care must be exercised when AI-assisted tools are used in connection with personal, confidential, judicial or child-related information.
SUGUD seeks to apply appropriate principles of:
data minimisation, purpose limitation, confidentiality, necessity and responsible information handling, taking into account applicable European and national data-protection requirements.
Sensitive information should not be introduced into an AI system merely because doing so is technically possible.
9. No Autonomous Decision-Making About Individuals
SUGUD does not use generative AI as an autonomous decision-maker for determining the legal rights, credibility, responsibility or alleged misconduct of identifiable individuals.
Human assessment and documentary verification remain essential to consequential conclusions concerning individuals or institutions.
10. Organizational AI-Literacy Measures
SUGUD’s AI-literacy framework may include:
- practical familiarisation with AI systems used by the organization;
- awareness of their capabilities, limitations and foreseeable risks;
- verification procedures for material outputs;
- enhanced review of legal and human-rights material;
- human oversight of consequential uses;
- periodic review of relevant AI legislation and regulatory guidance;
- safeguards proportionate to the context and risk of AI use; and
- documentation of relevant internal guidance, training or AI-literacy measures where appropriate.
These measures form part of SUGUD’s continuing organizational approach to responsible AI use and to supporting AI literacy in accordance with the applicable European regulatory framework.
11. Continuing Review
Artificial-intelligence technologies and their regulatory framework continue to develop.
SUGUD will therefore review this policy where appropriate in response to material technological developments, amendments to applicable legislation, regulatory guidance, changes in organizational practice or newly identified risks.
The organization may adapt its internal safeguards accordingly.
Accountability Declaration
Artificial Intelligence is a supporting instrument within SUGUD. It is not an evidentiary source, witness, legal authority or autonomous decision-maker.
The use of AI does not transfer responsibility for SUGUD publications, reports or submissions to an AI system or AI provider.
SUGUD retains human responsibility for its published material and seeks to ensure that material factual and legal assertions are traceable to verifiable sources and subject to appropriate human review.
The documented academic background of SUGUD’s founder in Information Systems, practical experience with digital and AI-assisted tools, human verification procedures and this organizational policy collectively form part of SUGUD’s approach to AI literacy and responsible AI use.
This policy forms part of SUGUD’s organizational measures concerning AI literacy under Article 4 of Regulation (EU) 2024/1689, as amended.
SUGUD – Advocacy for Families & Human Rights
AI Literacy & Responsible AI Policy
Effective: 30 September 2026