RESPONSIBLE COMPUTATIONAL IMAGING
Data protection and clinical accountability by design
AquaBIT intends to treat imaging data, reconstruction software and machine-learning models as safety-critical components—not as secondary digital features.
What AI may do
- Support artefact detection and signal-quality assessment.
- Assist physics-informed image reconstruction.
- Segment investigational conductivity maps.
- Quantify longitudinal changes for research comparison.
What AI may not do
- Replace clinical judgement or multidisciplinary review.
- Autonomously diagnose cancer or select treatment.
- Hide uncertainty, failed acquisition or missing data.
- Be trained or evaluated without appropriate governance.
Six governance commitments
Purpose limitation
Data will be collected for defined research, development or care purposes under an appropriate lawful basis and consent pathway.
Data minimisation
Only the information necessary for the stated purpose will be retained, with identifiers separated where practicable.
Security
Encryption, role-based access, audit logging, resilient backups and controlled data transfer will be required.
Independent evaluation
Models will be tested on data not used for training, with predefined metrics and external validation where appropriate.
Fairness
Performance will be examined across relevant body sizes, sex, age, tumour types and other clinically important groups.
Human accountability
Clinical responsibility will remain with appropriately qualified professionals within the intended workflow.
Research participants and future patients
People contributing data should receive understandable information about what will happen, why information is collected, who can access it, how long it will be kept and whether it may support future research. Withdrawal, deletion and data-sharing rights will be defined by the relevant consent, legal and regulatory framework.
Clinical integration
Any future connection with hospital systems will require documented interoperability, identity management, cybersecurity, information-governance review and clear separation between investigational output and validated clinical information.
Current position: AquaBIT does not currently operate a patient-facing imaging service or clinical AI system. These principles describe the intended governance framework for research and future development.
Discuss responsible data collaboration
We welcome expertise in medical imaging, clinical AI, cybersecurity, data governance and prospective validation.
AquaBIT Medical 