Dr Denis Newman-Griffis of the University of Sheffield offered a refreshingly human view of artificial intelligence. Rather than treating AI as a purely technical topic, they asked us to think about it as something that can connect and fragment people, information and systems. It is easy to talk about AI in terms of scaling and innovation. Dr Newman-Griffis reminded us that the real question is not simply what AI can do. The question is, ‘What does it do to the people who use it? To those affected by it? Or to those who are left out by it altogether?’
What can responsible practice look like?
Their career path has enabled them to reflect on AI from this interdisciplinary perspective. Beginning in computer science, moving into biomedical informatics and then social sciences, they now work at the intersection of AI and health. That combination seems notably important for responsible AI, because technology never exists in a vacuum. Human decisions and assumptions always shape it.
Dr Newman-Griffis described responsible AI as being effective, ethical and equitable. Responsible practice, they suggested, is more like washing with soap: a daily habit that's part of how we work, rather than a ‘special extra’.
As an example, they highlighted a project on AI and disability. Too often, digital systems reduce lived experience to narrow categories and miss important information that is not easily coded. In doing so, they risk losing the very human meaning behind the data. Dr Newman-Griffis described how natural language processing can be used to interpret text in ways that more faithfully capture lived experience.
Asking the right questions
In ethical AI, an important factor is asking the right questions. What we ask AI to do shapes what it produces, and who is affected (and in what way) by those outputs. We also need to trace design decisions carefully:
- Where does the data come from?
- What is included?
- What is missing?
- And how will the result actually be used?
These choices all exercise power.
The talk left me thinking that responsible AI is not only about better tools that describe data more fully, but about better judgement and transparency in how that judgement is made. We have active ethical choices to make in how we develop and use AI.
