Responsible Data Science by Design

At IDS, we undertake research on responsible data science by design. We look for technological innovations that can advance and enhance a responsibility agenda. For instance, how can individuals be given greater control over their data? It also means that we explore how varied responsibility agendas might influence innovation. For instance, how might privacy-preserving technologies differ across regions and countries with different ethical and legal approaches? 

Since its founding in 2017, IDS has pursued research topics that fall under “responsible data science by design,” such as advancing the FAIR principles. In 2020, we submitted a proposal for an international PhD training network to build a responsible automated discovery science ecosystem. All of these projects are undertaken on the principle that a responsibility agenda should not be “bolted on” to technological innovation but threaded throughout.  

Exciting research directions are opened up when we orient towards research about responsibility, rather than solely making recommendations about how to be responsible. 

What is responsible data science by design?

Responsible data science emerges when transparency, privacy/confidentiality, accountability, and social values are implemented and evaluated across:

  • actors (people and organisations) 
  • objects (data, algorithms, knowledge graphs, protocols) 
  • data science processes (procuring, analysing, and disseminating data), and 
  • social impacts (the intended and unintended consequences in the world).

"By design" refers to the need for:

  • ongoing ethical reflection
  • organizational commitments

  • responsive technology development, and
  • evaluative mechanisms.

We educate the next generation of responsible data scientists.

Through community events and an upcoming Master’s track in responsible data science, we aim to make sure that the next generation of data scientists can implement and evaluate the components of responsibility (transparency, privacy/confidentiality, accountability, and social values) in the relevant scope (across actors, digital objects, data science processes, and social impacts).

We help data-driven communities develop and implement a responsibility agenda that is grounded in data science.

There can’t be a one-size-fits-all responsibility agenda across every domain. Rather, communities should be entrusted with developing and implementing responsibility agendas that align with certain principles while matching their needs, goals, and histories. IDS is well-positioned to help, as we have experience with developing and implementing components of responsibility with multiple communities, and we have deep technical expertise in relevant problems. 

IDS was founded by Distinguished Professor Michel Dumontier, who is a co-founder of the FAIR (Findable, Accessible, Interoperable, Re-usable) principles. Our responsibility agenda is rooted in implementing FAIR workflows and disseminating FAIR practices across many domains. Those experiences have proven to us that the future of innovation lies in weaving a responsibility agenda with a technical agenda (not siloing them), and in creating future data scientists who are equally fluent in both agendas.