Human-centred AI is a guiding perspective in Don Sanjeewa Alwis’s work: people should be able to understand, question and influence the systems affecting them. Preparing for AI means considering its consequences for work, participation and human agency.
Keep oversight meaningful
A person asked to review an AI recommendation needs time, relevant information and authority to disagree. A confirmation button alone does not provide meaningful oversight. Decide who can interrupt a system and how affected people can challenge a result.
Build skills as responsibilities change
AI adoption changes tasks and expectations. Give people opportunities to learn the tools, recognise their limits and contribute to workflow design. Evaluate review workload and the experience of staff alongside output measures.
Ask who is missing
Language and access to resources shape participation in technology. Sanjeewa’s co-authored Sinhala AI research offers a concrete connection to this concern: adapting language models for a language with fewer digital resources.
Consider resources and consequences
Responsible adoption includes asking what computing resources a system needs, what value it creates and whether a simpler approach would meet the need. Sustainability is a consideration in design and evaluation, not a claim that every AI application has a positive environmental impact.