Frequently Asked Questions
General FAQs
- What is mpathic?
mpathic is a clinician-founded AI safety company focused on improving the quality and safety of human-AI conversations, particularly in high-risk moments like interactions with children, people with mental health needs, and in medical settings. We work with a large network of clinicians and safety experts and leverage our custom AI models to help organizations assess how AI systems perform in real-world interactions, identify risk, and close gaps where conversations fall short.
Our mission is to keep humans safe in the AI era through science-backed models, clinician expertise and high-quality datasets. - Why are clinicians relevant to AI safety?
In high-stakes, human-facing use cases, safety failures often show up in clinically meaningful signals such as distress cues, overconfident reassurance, dependency loops, or ungrounded clinical advice. Clinicians and behavioral experts help define what safe and effective interaction looks like in those contexts, and they help build evaluation frameworks that make findings repeatable and explainable. - What is mpathic Studio?
mpathic Studio is a proprietary application designed to meet the surging demand for clinically grounded AI safety infrastructure. With mpathic Studio, experts and customers are able to test model behavior, monitor live interactions, and ensure AI-generated responses meet the highest standards of clinical accuracy and risk awareness. - How do you select mpathic experts?
Our experts are hand-selected, licensed psychiatrists, doctors, clinicians, mental-health and other safety professionals. Each expert goes through extensive credential and background checks, interviews and aptitude tests to ensure the highest level of quality and expertise. Learn about becoming an expert here. - Can mpathic be added to existing workflows
Yes, we can integrate our products and services into existing workflows. mpathic Studio offers API integration, dashboards, and unified workflows, or larger clients may choose to embed mpathic experts into their existing workflows.
Data Privacy, Safety and Security FAQs
- How does mpathic prioritize individual privacy?
At mpathic, privacy is highly esteemed. Extensive measures are implemented to ensure the confidentiality and protection of personal data. Strict privacy safeguards are maintained to uphold top-tier security and privacy. A comprehensive privacy policy is available with detailed information regarding the collection, use, sharing, and processing of personal data. For an in-depth grasp of privacy practices, refer to the Privacy Policy section. - What measures does mpathic take for data security?
Data security at mpathic is primarily shaped by adherence to international data protection laws and regulations, such as the EU and UK General Data Protection Regulation (GDPR). A strong emphasis is placed on safeguarding personal data, and the pride in SOC 2 Type II compliance reflects a commitment to the highest standards of security, confidentiality, and data integrity. More information about data security can be found in the Information Security section. - What type of testing does mpathic perform to ensure data security?
To certify the robustness and effectiveness of data security measures, annual penetration testing is conducted by an independent third party. - How does mpathic ensure HIPAA compliance?
To ensure understanding of privacy and security requirements applicable to protected health information, training is provided during the onboarding process. Any HIPAA-covered PHI is processed using HIPAA-compliant tools selected through a meticulous vendor selection process. - Does mpathic meet SOC 2 Type II standards?
Yes, all sensitive information handled by mpathic is managed in full compliance with SOC 2 Type II standards. - How does mpathic ensure GDPR compliance?
Personal data covered by GDPR is processed using GDPR-compliant tools, which are chosen following a comprehensive vendor selection process. - What additional data security measures does mpathic take?
Techniques are employed to segment data when creating custom models, ensuring the data remains segregated and secure as explicitly outlined in the data use agreement. In addition, there are dedicated and isolated hosting options for enhanced data security. - What happens with my data when we integrate mpathic API?
mpathic processes your data but does not automatically store your data unless requested and agreed upon in writing. Sensitive and personal data is processed using services within Amazon Web Services (AWS), and various approved tools. All data within these systems is encrypted at rest and in transit to ensure it’s safe, secure, and available only to intended recipients. HTTPS encryption is configured for customer web application access. - Does mpathic use customer data to train its models?
Before training models with customer data, mpathic obtains prior written consent. Any use of customer data is specified in the data use agreement before implementation. - Is there an option to anonymize or pseudonymize sensitive data?
mpathic provides an option for customers to pseudonymize or anonymize sensitive data. We can also manually label for personally identifiable information (PII) if required. If you have specific requirements to pseudonymize sensitive data, contact us and we’ll work with you! - Who has access to customer data and the outputs?
mpathic implements role-based access controls for information system access. Access controls to sensitive data in databases, systems, and environments are granted on a need-to-know/least privilege necessary basis. mpathic employees with special job functions may have access to certain additional resources. - How does mpathic select vendor partners and business associates?
mpathic is HIPAA, GDPR, and SOC 2 Type II compliant, so any vendors we work with that collect or process sensitive data must go through a critical vendor selection process. They are also required, pursuant to the terms of their agreements with us, to implement required safeguards. Additionally, AWS and Google, the web service providers that we contracted, are both HIPAA-compliant.
AI Medical Monitoring FAQs
- What is AI-augmented medical monitoring, and how does it work?
AI-augmented medical monitoring combines natural language processing (NLP) and advanced analytics to analyze clinical trial data alongside human medical monitoring. It utilizes technologies like mTrial to review audio or transcriptions, ensuring data and site consistency while monitoring the quality and adherence of therapeutic-assisted trial delivery. The AI performs an initial pass on reviewing the data, providing insights and flagging moments for further review and confirmation by humans. - How can AI medical monitoring improve patient safety and trial integrity?
AI medical monitoring contributes to better patient safety and overall trial integrity by enabling real-time monitoring and early detection of potential safety concerns. It provides timely alerts and insights that help prevent adverse events, enhance patient safety, and ensure the integrity of trial data. Additionally, AI technology allows for the comprehensive review of audio recordings, enabling precise identification of data patterns and accurate analysis with a higher level of accuracy compared to human medical monitors. - Is AI medical monitoring compliant with regulatory standards such as HIPAA and GDPR
Yes, AI medical monitoring solutions like mTrial are compliant with regulatory standards such as HIPAA and GDPR. Compliance with regulations is crucial for all stakeholders involved in clinical trials, ensuring data privacy and security while adhering to legal requirements. - How does AI medical monitoring handle patient data privacy and security?
AI medical monitoring solutions prioritize patient data privacy and security by utilizing state-of-the-art encryption methods and adhering to regulatory standards such as HIPAA and GDPR. Data is handled in a secure manner, with measures in place to prevent unauthorized access or breaches. - Can AI medical monitoring be customized to meet specific trial requirements?
Yes, AI medical monitoring solutions offer customization options to meet specific trial requirements. Whether it’s adjusting algorithms to focus on particular clinical behaviors or tailoring metrics to match desired outcomes, customization ensures the technology aligns with the unique needs of each trial. - What assurances do we have regarding the accuracy of AI-generated insights?
AI medical monitoring solutions provide assurances regarding the accuracy of AI-generated insights through rigorous testing, validation processes, and ongoing refinement. Accuracy levels are continuously monitored and fine-tuned to ensure reliable and precise results. - Does AI medical monitoring replace the need for human medical monitors?
No, AI medical monitoring does not replace the need for human medical monitors. While AI augments decision-making capabilities and enhances efficiency, expert medical decisions are still required to be made by trained humans. AI serves as a supportive tool, increasing the reach and efficiency of medical monitors. - How does AI medical monitoring contribute to cost savings and trial efficiency?
AI medical monitoring contributes to cost savings and trial efficiency by streamlining data analysis processes, reducing manual review times, and providing insights that lead to faster decision-making. The technology improves trial fidelity and safety while saving time and resources. - Can AI medical monitoring integrate with existing systems or workflows?
Yes, AI medical monitoring solutions can integrate with existing systems or workflows, providing seamless compatibility and enhancing operational efficiency. Integration options ensure smooth adoption and utilization within established processes. - What support and training options are available for users of AI medical monitoring technology?
Users of AI medical monitoring technology have access to comprehensive support and training options. This includes onboarding assistance, ongoing technical support, and training sessions tailored to the specific needs of users. The goal is to provide users with the knowledge and resources to effectively utilize the technology and optimize trial outcomes.