Can we understand every line of an AI algorithm to understand a decision made with the use of artificial intelligence?
The answer is no. In many cases, such an approach would be neither possible nor genuinely useful.
In the Burn The Map podcast, Gabriela Bar discusses what truly matters when AI systems influence decisions affecting individuals. The key question is not whether every technical detail of an algorithm can be explained, but whether a person affected by an AI-supported decision can understand:
- why the decision was made,
- which factors had the greatest impact,
- what steps can be taken to challenge the decision or seek a different outcome.
As Gabriela Bar explains, with increasingly complex AI models, full technical transparency may be difficult to achieve. However, this does not reduce the responsibility of organisations to provide individuals with clear, understandable and meaningful information – particularly when AI influences access to public services, benefits or other important opportunities.
Explainability should not be limited to describing a model’s technical architecture. Its purpose is to help people understand the consequences of AI-assisted decisions and enable them to take informed action.
During the conversation, Gabriela Bar also addressed broader issues related to AI Law, responsible AI governance and the role of transparency in building trust in artificial intelligence systems.
She thanks Dan Baird for the invitation and for highlighting important discussions around the legal and ethical challenges connected with AI.
🎧 The full conversation is available on YouTube and Spotify:
YouTube: Gabriela Bar – Burn The Map Full Episode
Spotify: Burn The Map via Spotify
