At the latest TechEx North America event we spoke with Ilana Golbin Blumenfeld, Partner at PwC, about the complex nature of bias in AI systems. We explored how bias can arise not only at the point where an AI model produces results, but throughout its entire lifecycle. This includes the way data is collected, the populations it reflects, the objectives it is designed to serve, and how it is ultimately used in decision-making processes.

We also examined how organisations can choose between open-source and proprietary AI models, taking into account considerations such as transparency, performance, and operational feasibility. Ilana highlighted the importance of defining success criteria early, selecting appropriate technology accordingly, and ensuring that teams are properly trained to use AI responsibly. We discussed the role of governance frameworks and procedural safeguards, as well as the importance of understanding AI as a component within a broader system rather than a standalone solution. 

Check out the full interview here: