Responsible AI Principles

Updated

Website: https://zweigadvisory.com

Joint Legal Entities ("Zweig Advisory," "we," "us," "our"): This policy governs the development, deployment, and use of Artificial Intelligence (AI) systems across both entities.

  • Zweig Services and Asset Management GmbH (German/EU Entity)
  • Zweig Cost Control & Risk Management Services EST. (UAE Mainland Entity)

Vision and Ethical Commitment

Zweig Advisory recognizes the transformative potential of Artificial Intelligence to enhance productivity and deliver strategic value to clients. We are committed to developing and deploying AI systems in a manner that is trustworthy, safe, and aligned with human rights, democratic values, and principles of sustainable growth.

Our Responsible AI Principles are founded on global best practices, including the OECD AI Principles, and rigorously apply the requirements of the European Union, Germany, and the United Arab Emirates.

1. Core Responsible AI Principles

We adhere to the following values-based principles in every stage of the AI system lifecycle:

Principle Commitment Regulatory Relevance
Human Oversight & Accountability AI systems must be subject to appropriate human oversight and clear accountability mechanisms. Responsibility for the outcomes of AI-driven decisions always rests with a designated human. EU AI Act, GDPR
Fairness & Non-Discrimination We commit to designing and training AI systems to minimize the risk of discriminatory outcomes and prohibited bias, particularly in high-stakes areas like financial risk assessment. EU AI Act, OECD Principles
Transparency & Explainability AI systems must be understandable. We commit to informing stakeholders when they are interacting with an AI system and to providing sufficient explanation of AI-driven decisions, especially those impacting individuals' rights. EU AI Act, OECD Principles
Robustness, Security & Safety AI systems must be technically robust, reliable, and secure against physical, cyber, and adversarial attacks. They must maintain appropriate levels of accuracy and operate safely. EU AI Act, OECD Principles
Data Protection & Privacy AI development must prioritize the protection of Personal Data, ensuring strict compliance with all applicable privacy frameworks, including the GDPR and UAE PDPL. GDPR, UAE PDPL

2. Governance and Regulatory Compliance

Our framework is designed to manage the complexity of operating under the world's most stringent AI and data laws:

A. EU AI Act and German Alignment (Zweig Services and Asset Management GmbH)

Our German entity ensures that any AI system developed or deployed for clients within the EU meets the stringent requirements of the EU AI Act, particularly for "High-Risk" applications relevant to financial advisory, such as creditworthiness assessment, risk scoring, and internal risk modeling. This includes:

  • Risk Mitigation: Conducting adequate risk assessment and implementing mitigation systems before deployment.
  • Data Quality: Ensuring high quality and representative datasets are used to train the system to minimize the risk of bias.
  • Traceability: Maintaining detailed documentation and logging of all activity to ensure the traceability of results and allow for post-deployment monitoring.
  • Human Oversight: Designing the system to allow for human intervention and ultimate responsibility for the final output.
  • Cybersecurity: Ensuring a high level of cybersecurity and resilience appropriate for the risk profile of the application.

B. UAE National Strategy Alignment (Zweig Cost Control & Risk Management Services EST.)

We align with the visionary goals and governance structures of the UAE's National Strategy for Artificial Intelligence. Our AI activities in the UAE Mainland adhere to the ethical guidance published by the UAE government (e.g., the AI Ethics Guide), emphasizing efficient, secure, and socially responsible innovation. We actively participate in the development of responsible AI practices within the region, ensuring that local expertise guides our ethical implementation.

3. Application and Risk Management in Consulting

We recognize that AI can generate specific risks in the financial and consulting sector, including systemic risk, market concentration, and misplaced dependence. Our approach to managing these risks includes:

Proprietary Model Transparency: We strive to overcome the "black box" challenge inherent in complex proprietary algorithms by documenting model inputs, constraints, and decision logic to provide clients with necessary transparency and confidence.

Bias Mitigation in Datasets: We rigorously audit data sources used for AI training to detect and correct historical or systemic biases that could lead to unfair outcomes in client assessment or resource allocation.

External Dependency Management: We assess the concentration risk associated with using major third-party AI suppliers, ensuring that technology dependence does not increase operational or systemic risk for our firm or our clients.

Incident Reporting: We commit to tracking, documenting, and reporting serious incidents or systemic risks arising from our AI systems to relevant regulatory authorities without undue delay, in line with emerging EU and global requirements.

4. Training and Oversight

Our commitment to Responsible AI is enforced through mandatory internal training and clear lines of accountability:

Training: Specialized training is provided to employees involved in the design, procurement, and deployment of AI systems, focusing on detecting bias, ensuring data governance, and implementing human oversight measures.

Governance: The central Compliance function maintains oversight of the AI governance framework, ensuring adherence to the principles and conducting regular audits of deployed AI models.

Approval

This Responsible AI Principles Statement has been approved by the Senior Management of Zweig Advisory, demonstrating the Executive Team's commitment to ethical AI deployment.

Signed by:

Name: Margarita Zweig, Alexander Zweig
Title: Managing Directors

5. Contact and Entities

For inquiries regarding our AI Governance and ethical compliance, please contact the Compliance Department.

Joint Legal Entities and Registered Addresses:

Entity Registered Address Primary Jurisdiction
Zweig Services and Asset Management GmbH Potsdamer Straße 92, 10785 Berlin, Germany EU/Germany
Zweig Cost Control & Risk Management Services EST. Westburry Tower 1, Business Bay, Dubai | UAE UAE Mainland

Central Email Contacts: