Artificial Intelligence Policy

1. Data Privacy & Security 

• Client Data Usage: Meridia strictly prohibits the use of client data for training or testing generic AI models. In instances where machine learning is employed to improve the performance of models within the scope of a specific project, explicit client consent is obtained, and only the client's own data is used. This learning is not shared between clients, ensuring the confidentiality and integrity of each client's information.
• Data Access: Meridia adheres to the principle of least privilege, granting access to client data only to authorized personnel and solely for the purpose of providing the agreed-upon services. We have strict internal policies and technical controls in place to prevent unauthorized access or misuse of client data.
• Data Sharing: Meridia never shares client data with any third parties, including public AI models. All AI models utilized within the Meridia platform are private versions hosted securely within our own environment, isolated from external systems.
• Data Logging and Monitoring: Access to client data is meticulously logged, providing a comprehensive audit trail for accountability and security purposes. Our system automatically monitors access patterns and user behavior to detect any anomalies or suspicious activities. Any identified issues are promptly investigated and addressed by our security team.
• Data Encryption and Protection: All client data is encrypted at rest and in transit using industry-standard encryption algorithms. We employ a multi-cloud strategy to ensure data redundancy and resilience. Regular backups are performed, and the system fully version-controls data, allowing for quick recovery and rollback to any point in time in the event of an incident.

2. AI Model Management 

• Cloud Environments: Meridia leverages the capabilities of multiple cloud providers, with Microsoft Azure as our primary environment. We also utilize AWS and Google Cloud to take advantage of their latest technologies and AI models, ensuring a robust and diversified infrastructure
• Supported AI Models: Our platform offers multi-model access, including full support for industry-leading AI models such as ChatGPT, Claude, Gemini, and Lamma. This enables us to provide a wide range of AI-powered functionalities to our clients. We continuously evaluate and integrate new models as they become available to stay at the forefront of AI advancements.
• Model Selection and Optimization: The Meridia platform employs intelligent algorithms to select the best AI model for each specific task, such as information summarization, narration, image recognition, and logical reasoning. It seamlessly transitions between models as engagements progress to ensure optimal performance and accuracy. We continuously fine-tune and optimize our model selection process based on performance metrics and client feedback.
• User Control and Customization: We empower our users with control over their AI interactions. Users can request the AI assistant, George, to retry a task, provide additional context, or add nuance to the outputs. For clients with specific requirements, we offer the ability to select preferred models or customize the AI behavior to align with their unique needs.
• Model Updates and Maintenance: Meridia regularly updates the AI models integrated into our platform to ensure access to the latest and most capable versions available. We have a dedicated team responsible for monitoring model performance, applying security patches, and conducting routine maintenance to ensure optimal functionality and security.

3. Data Governance & Compliance 

• Secure Authentication: Meridia integrates seamlessly with clients' existing Office365 and Azure Entra ID tenancy for secure authentication. This allows clients to leverage their own identity and access management systems, ensuring consistent and robust authentication processes.
• Model Independence and Diversity: Meridia leverages the diverse capabilities of the AI landscape and does not rely on any single model or provider. By integrating with multiple models and maintaining a multi-model ecosystem, we mitigate risks associated with dependence on a single AI technology and ensure the availability of a wide range of capabilities.
• Rapid Model Integration: Our platform is designed to quickly integrate new AI models as they become available. We have streamlined processes in place to evaluate, test, and deploy new models within hours, ensuring our clients have access to the latest advancements in AI technology.
• Data Usage and Training: Meridia strictly prohibits the use of client data for training AI models. We do not use any client information to improve or enhance the performance of our AI models. Our models are trained using carefully curated and vetted datasets that comply with legal and ethical standards.
• Regulatory Compliance: Meridia operates in compliance with relevant data protection regulations, such as the General Data Protection Regulation (GDPR) and industry-specific guidelines. We regularly review and update our policies and practices to ensure ongoing compliance with evolving regulatory requirement.

4. Transparency & Accountability 

• Transparency: Meridia is committed to being transparent about our AI usage practices. We provide clear information to our clients about how AI is utilized within our platform, including the types of models used, data handling practices, and security measures in place. We encourage open communication and are ready to address any questions or concerns clients may have.
• Accountability: We hold ourselves accountable for the responsible and ethical use of AI technologies. Our executive leadership team oversees the implementation and enforcement of this AI usage policy. We have designated roles, such as the Data Protection Officer and AI Ethics Officer, who are responsible for ensuring compliance and addressing any issues that may arise.
• Audits and Assessments: Meridia regularly conducts internal audits and engages independent third parties to assess our AI usage practices. These assessments help us identify areas for improvement, validate our compliance with industry standards and regulations, and provide assurance to our clients regarding the integrity of our AI systems.
• Incident Response and Reporting: In the event of any AI-related incidents or breaches, Meridia has a comprehensive incident response plan in place. We promptly investigate and address any issues, and we notify affected clients in a timely manner. We maintain detailed incident logs and conduct post-incident reviews to identify root causes and implement necessary remediation measures.

5. Continuous Improvement 

• Stakeholder Engagement: Meridia actively engages with our clients, industry partners, and other stakeholders to gather feedback, insights, and best practices related to AI usage. We value the input of our clients and incorporate their suggestions and concerns into our AI development and deployment processes.
• Research and Innovation: We dedicate resources to research and innovation in the field of AI, staying up-to-date with the latest advancements and exploring new ways to responsibly leverage AI technologies for the benefit of our clients. We collaborate with academic institutions, research organizations, and industry experts to contribute to the broader AI community and drive responsible AI practices.
• Employee Training and Awareness: Meridia invests in the training and education of our employees to ensure they have a deep understanding of AI technologies, ethical considerations, and best practices for responsible AI usage. We conduct regular training sessions, workshops, and awareness campaigns to foster a culture of AI responsibility and accountability throughout our organization.
• Policy Review and Update: This AI usage policy is a living document that is regularly reviewed and updated to reflect advancements in AI technologies, changes in regulatory landscapes, and evolving industry best practices. We encourage our clients and stakeholders to provide feedback and suggestions for improvement. Any updates to the policy will be communicated to all relevant parties and made easily accessible.

Contact us