AI Guidelines Policy for the Silat Wassel Platform

Issued: 14/4/2026

This policy outlines the ethical and operational framework for the use of artificial intelligence (AI) on the Silat Wassel platform. We are committed to using AI responsibly to foster communication while upholding the values of human dignity, trusted resources, and social equity. This policy is a living document and is continually updated to reflect technological advancements and emerging regulatory landscapes.

1. Our Definition and Approach to AI

We view AI not merely as a tool but as an interdisciplinary research field that replicates human intelligence via computational technology. Our approach is guided by:

Human-Centered Design & Hope-Based Communication

We prioritize a human and hope-based approach to all digital interactions. AI is utilized to bridge gaps in communication, ensuring that technology serves the dignity and aspirations of our audience rather than replacing the human element.

Critical Awareness of Socio-Political Context

We recognize that AI is not a neutral tool; it is deeply embedded in socio-political contexts. Because AI is trained on data that often represents dominant versions of the world, it can reproduce existing social biases. We remain critically aware of these inherent limits and the potential for harm in data representation.

Response-Ability & Human Oversight

We are committed to creating “ecosystems of response-ability.” This means that we, as humans, hold ultimate responsibility for the outcomes, regulations, and effects of AI interactions. AI serves as an assistant for decision-making, but it does not carry final authority.

Operational Workflow: From Generation to Approval

To ensure consistency across all teams, the following workflow is mandatory for all AI-assisted tasks:

1. Drafting/Generation: Utilize the “Context + Instruction” model to prompt the AI.
2. Internal Review: Initial check by the creator for tone, mission alignment, and obvious errors.
3. Verification: Cross-referencing outputs with the “Trusted Resource Mapping” (see Section 4).
4. Final Approval: A designated team lead must review and approve the output before it is published or implemented.

2. Risk-Based Categorization and Oversight

We distinguish between different types of AI use to ensure that higher-risk applications receive appropriate scrutiny:

3. Data Integrity and Bias Mitigation

Data is a symbolic abstraction of reality. To prevent the reproduction of biases, Silat Wassel implements the following safeguards:

• Bias Identification: We actively monitor for biases in allocation (e.g., gender-based selection) and service. Teams must explicitly prompt the AI to draw from diverse perspectives and look for gaps in the representation of subpopulations.
• Representation: We strive to ensure “human” is not defined by a narrow demographic.
• Dignity: We strictly prohibit the generation of content that makes offensive generalizations or dehumanizing comparisons.
• Flagging Mechanism: If a team identifies a persistent bias in a tool’s output, they must escalate the issue to the AI expert to adjust the prompting strategy or tool selection.
  • Bias and Representation Check:
    • Are different groups or perspectives fairly represented?
    • Is any group portrayed in a stereotypical or narrow way?
    • Has the AI defaulted to a dominant or Western-centric perspective?
    • Have we prompted or validated against diverse sources where relevant?

4. Fact-Checking & Trusted Resources

To combat the risks of AI “hallucinations,” Silat Wassel implements a strict verification layer:

• Manual Verification: All AI-generated content intended for public consumption must be reviewed by the team to ensure accuracy.
• Trusted Resource Mapping: Our AI systems are directed to prioritize information from civil society manifestos, organized professional networks, and local partner data.
• Step-by-Step Verification Workflow:
  1. Identify key claims, statistics, or citations in the AI output.
  2. Verify each claim against a “Human-Verified” source or a trusted organizational database.
  3. If a claim cannot be verified through a trusted source, it must be removed.

5. The Silat Wassel AI Expert

We have developed a specialized AI expert to support both our team and our audience. This expert is trained to:

• Address Inquiries: Take any questions or suggestions to help improve platform operations.
• Safe Interaction: Operates under a “Context + Instruction” model to provide specific, high-quality guidance.
• Consultation: Teams must consult the AI expert for all “high-risk” tasks or when troubleshooting complex ethical dilemmas.
• Privacy First: Users are reminded never to share sensitive personal information with this or any AI tool.

You can reach out to our AI expert contact via:

Email: media@silatwassel.com or WhatsApp Phone Number: +20 10 40598238

6. Generative AI and Prompting Strategy

When using Generative AI (Gen AI), we prioritize the following:

• Specificity over Randomness: Using detailed prompting strategies to ensure outputs align with our mission.
• Plausibility Checks: Recognizing that Large Language Models (LLMs) predict patterns to generate “plausible” responses, which must always be fact-checked by our human team.

7. Accountability and Governance

Silat Wassel adheres to the principles of the Civil Society Manifesto on AI.

• Decision Support: AI should assist in making decisions, not replace human judgment.
• Democratization: We involve our diverse audience in the conversation regarding how these tools are used.
• Living Document Governance:
  • Review Schedule: This policy is reviewed and updated every six months.
  • Responsibility: The Silat Wassel administrative team and the AI expert are responsible for maintaining and updating these guidelines.
  • Feedback Loop: Teams are required to submit monthly feedback on AI tool performance to support continuous improvement.