Daiser Transparent AI Ethics Framework

At Daiser, we build world-class Software as a Medical Device using the latest AI and agentic AI. Because our technology supports health and care decisions at the moments that matter most, we hold our AI to a higher standard. 

This framework sets out the principles that govern how AI is designed, deployed, and monitored across the Daiser platform, including Daiser Proof, Guide, and Care.

This framework applies to every AI system we build or configure, from conversational health tools and clinical decision support to agentic AI workflows and virtual care teams.

1. Safety First, Always

Our AI operates within regulated guardrails, developed under ISO 13485 and IEC 62304 and aligned to the relevant medical device classification (Class I and Class II). AI capabilities are only deployed where their safety and performance have been verified for their intended use. Where AI informs clinical decisions, it does so as a support to — never a replacement for — appropriate clinical judgement.

2. Human Oversight and Accountability

Every agentic AI workflow on the Daiser platform is designed with human oversight. Care pathways, escalation rules, and agent behaviours are defined, reviewed, and approved by qualified people before deployment. Accountability for outcomes always rests with identifiable humans and organisations, never with an algorithm. Users and partner organisations can review, override, and switch off AI-driven functions.

3. Transparency

People have the right to know when they are interacting with AI. Our conversational and agentic tools identify themselves as AI, and we document (in plain language) what each AI system does, the data it uses, and its known limitations. Partner organisations receive the technical and regulatory documentation needed to understand and explain the AI they deploy.

4. Privacy and Data Stewardship

AI on the Daiser platform is built on the same foundation as the rest of our infrastructure: ISO 27001-certified information security and strict adherence to applicable data protection law, including UK GDPR. Health data is processed only for defined, lawful purposes. We do not use personal health data to train AI models without an appropriate legal basis and, where required, explicit consent. See our [Privacy Policy] for full details.

5. Fairness and Equity

Better health and care should reach everyone. We assess our AI systems for bias across populations, monitor real-world performance for unequal outcomes, and design for accessibility and inclusion. Where we identify disparity in how an AI system performs, we act to correct it.

6. Clinical and Scientific Integrity

Our AI is grounded in evidence. Models and decision-support logic are built from validated clinical sources, and claims about performance are supported by data generated within our regulated research environment. We are designers, AI engineers, and scientists in equal measure, and we do not deploy AI whose clinical basis we cannot defend.

7. Continuous Monitoring and Improvement

AI in healthcare is never done. We continuously monitor deployed AI for accuracy, drift, safety signals, and unintended behaviour, under our post-market surveillance obligations as a medical device manufacturer. Issues are logged, investigated, and remediated through our quality management system, and material changes are re-verified before release.

8. Governance

This framework is owned by Daiser's leadership and overseen by our AI governance function, which includes regulatory, clinical, and engineering representation. It is reviewed at least annually and whenever there is a significant change in technology, regulation, or the way our platform is used. All Daiser employees and partners configuring AI on our platform are required to work within it.