Keynotes

We often hear that AI systems are biased, but what does that actually mean? While it is clear that discriminatory or unfair AI systems can erode public trust and hinder adoption, the nuances of bias are not easily understood. In this talk, Andrew will discuss the rhetoric around bias and challenge the oversimplified notion of bias as a purely negative phenomenon. He will delve into the root causes of AI bias, examining how these may be introduced into AI systems and how we can conceptualise these in human terms. By changing the language that we use to talk about bias, we can move beyond objectives to mathematically eliminate bias, but to build towards fairness and safety instead. We can evaluate whether the objective functions of generalist AI systems are appropriate for the specific use cases we envisage. This talk aims to foster a critical conversation about AI bias, moving beyond simplistic narratives and towards a more comprehensive approach towards addressing this complex issue.

In an era where technology continually reshapes research landscapes, the role of research software is pivotal in driving scientific progress and societal impact. This presentation showcases how a regional research institute, outside major townships, is making a genuine impact on world-class research while maintaining a connection with the community.

It is often acknowledged that research and its impact are not always synonymous. In the area of medical research, it takes time for research to make an impact in the community – and direct benefits to the participants are rare. Adding to this is the lack of data standards between different research communities, and other research barriers.

To address this disconnect, better engagement is a solution – research engagement that brings the community alongside throughout the process, with timely reciprocity, fosters a better understanding of the research. This mutual understanding and trust are essential for genuine relationships. The talk will highlight the role of Research Software Engineering (RSE) in connecting innovation with tangible outcomes for the local community, particularly in addressing real-world challenges such as accessibility and underrepresentation.

For the research communities around Aotearoa New Zealand, the collaborative discussions and interdisciplinary partnerships central to RSE are crucial for maximising the impact of software innovations. Our attempt to keep core values for research in check by embedding CARE principles into our operational practices will be outlined, while FAIR principles help develop robust tools and maximise the value of collected data. This abstract has been co-authored by Eryn Kwon, Joshua McGeown, Maryam Tayebi, Leigh Potter, Samantha Holdsworth from the Mātai Medical Research Institute.