Trust
Ethics & Responsible AI
Artificial intelligence can create important opportunities in healthcare - but only when it is developed responsibly. Cognitive HealthTech is committed to developing AI and behavioural-analysis technology in a way that respects people, supports healthcare professionals and recognises the limitations of emerging technology.
Human-centred by design
Technology should support people rather than reduce them to a score or data point. Our approach begins with the individual and considers cognitive change in the context of their own patterns over time.
Supporting, not replacing, clinical judgement
CHT is being developed to provide additional information and context. It is not intended to replace appropriately qualified healthcare professionals or clinical judgement.
Our current technology is progressing through development and validation and should not be interpreted as providing an independent medical diagnosis.
Responsible use of AI
Our development principles include:
- Transparency
- People should have meaningful information about when and why AI is being used.
- Human oversight
- Important healthcare decisions should remain appropriately human-led.
- Privacy
- Behavioural and health-related information requires particularly careful handling.
- Fairness
- We aim to understand and reduce inappropriate bias rather than assume AI is inherently unbiased.
- Validation
- Performance claims should be supported by appropriate evidence.
- Proportionality
- Technology should collect and process information appropriate to the purpose for which it is being used.
Accountability
Responsibility for technology cannot simply be delegated to an algorithm.
Personal baselines and fairness
One area of particular interest for CHT is the use of an individual's own behavioural patterns as context for understanding change.
This approach is being explored as a way to complement traditional population-level comparisons and potentially provide a more personalised picture of cognitive change.
It does not remove the need for rigorous scientific and clinical validation.
No hidden surveillance
CHT is intended to analyse information within transparent, appropriate and consented interactions. We do not believe cognitive-health technology should depend on covert monitoring.
Continual review
Responsible AI is not a one-time certification. As the technology, evidence base and regulatory environment evolve, CHT will continue to review its approach with input from researchers, clinicians, people with lived experience and relevant specialists.