AI ethics and safety

Developing responsible AI

We collaborate with researchers, academics, clinicians, and patient advocacy groups to advance responsible and safe AI that delivers patient and societal benefit.
Our approach

Harnessing AI for a positive impact

Artificial Intelligence can provide extraordinary benefits, but like all technology, it can have harmful impacts unless it is designed, developed, validated, and deployed responsibly. At Owkin, we think deeply about the ways in which we build AI that benefits society while working to reduce bias and unfairness. 

As we incorporate cutting-edge technologies like generative AI and foundation models, we are developing tools and processes to assess the impact these new ways of generating knowledge will have. Our aim is to engineer technologies and systems that achieve positive outcomes for patients and broader society.

Our focus

Healthcare is one of the most exciting fields for AI research. It is also one of the most sensitive and highly-regulated industries in the world. At Owkin, we have pioneered privacy-enhancing technologies, like federated learning that protect data rights and privacy while also unlocking AI’s capabilities to discover and develop better treatments for patients. Our mission has grown to apply our privacy enhancing technologies to the real world challenges facing drug discovery, development, and diagnostics.

Our AI safety & ethics team focuses on: 

Mindful deployment of AI into complex healthcare systems.

Emerging threats
Emerging threats

Emerging threats to patient privacy and ways to mitigate against it.

Minimize bias
Minimizing bias

Decision making during model building - how do we decide what is and isn’t included in the model and how do we ensure bias is minimized?


Validation of AI models that account for heterogeneity and brings humans back in the loop.

Decision making
Decision making

Algorithmically enhanced decision making across diagnosis, care, and drug discovery and development.

Our approach

Co-designing ethical and safe AI

At Owkin, we co-design our AI with researchers and academics to ensure that we conceptualize and implement tools and techniques to help us anticipate ethical harms, so these can be proactively mitigated. We include patients and clinicians as true partners in the development and deployment process.

We recognize that not all ethical harms can be predicted in advance, society's expectations change and harms can be latent. Instead we focus on:

  • anticipating potential harms
  • taking a reflexive and pro-ethical approach to design
  • ensuring we are regularly re-evaluating our products

We are focused on training and developing our teams with the skills to think pro-ethically and create standardized, but practical, processes for embedding ethical thinking into AI development. 

Contact us