Our AI Principles

At mpathic, we understand the transformative power of technology, specifically AI, and we’re committed to harnessing this power with integrity and care. Our AI principles are not merely a framework; they serve as a testament to our unyielding commitment to advancing technology in ways that are conscientious, considerate, and respectful. Central to our principles is the concept of empathy—that our AI technology should consistently enhance human capabilities, respect user privacy, and embody impartiality, without causing harm.

These principles guide us in our ongoing mission to create AI solutions that earn our users’ trust, exceed their expectations, and acknowledge the critical importance of empathy in all interactions. They form the backbone of our approach to the research, design, development, and deployment of AI technologies, shaping our pursuit of AI that does more than simply innovate—it strives to improve the world we live in.

Principle	Commitment
Practice integrity in every interaction.	Integrity of an AI system assures that the model treats people with consistent and impartial treatment across all scenarios.
Build robust systems.	Robustness of an AI system verifies the model’s accuracy across all potential use cases while showcasing resilience against malicious attacks.
Foster transparency and explainability in all decisions and actions.	Transparency of an AI system refers to the ability to explain and replicate the decisions made by the system. To achieve transparency, predictions should be accessible, understandable, and occur promptly for users, developers, and all other key stakeholders.
Incorporate privacy in all handling of personal information.	Privacy in an AI system relates to protecting sensitive information used in training, validating, testing, and/or in the ongoing use of the model. An AI system should be designed so that it cannot divulge private information about individuals in the training data, nor should it be manipulable to reveal sensitive information about a person through malicious inputs like “jailbreaking” an LLM-based chat.
Strive for accountability in all actions.	Accountability of an AI system indicates the readiness of the system’s designers or operators to respond to feedback and appeals and to put remediation mechanisms in place when issues arise. Operators can explain system results and decision, if their functionality and decisions can be explained, to users, governing bodies, and other key stakeholders to ensure compliance with laws and ethical standards.
Prioritize safety in all environments and situations.	The safety of an AI system emphasizes that it does not cause harm to individuals, environments, or societies. This includes preventing the deployment of an AI system in appropriate contexts and/or among populations for which it was not adequately designed. This could occur when the AI system has not been meticulously calibrated to be aligned with clinical and/or legal standards.