What happens when bugs don’t just cause a crash, but destroy human lives? In a world that increasingly delegates critical decisions to algorithms, the line between “statistical optimization” and “systemic injustice” has become dangerously thin. In this talk, we’ll explore the fine line between cognitive error and computational bias, analyzing how justice is sliding toward opaque automation.
The keynote address:
- Human Error: We’ll analyze the cognitive biases that affect real-world courts. Through studies and documented news stories, we’ll see how environmental factors and systemic biases influence the decisions of human judges.
- Machine Error: We’ll sift through real-world cases of algorithmic discrimination. We’ll see where AI has failed (from the COMPAS case to credit scoring and surveillance systems), deconstructing the intrinsic problems of machine learning: from the toxicity of datasets to the lack of explainability.
- The Dystopian Warning: We will use an emblematic case to illustrate a risk as dangerous as it is subtle: the loss of the right to appeal and empathy. What happens when AI ceases to be a support tool and becomes an absolute, final and context-free judge?
Why watch this talk: This talk invites reflection on the structural responsibility of those who write code. The goal is not to demonize technology, but to understand that in an “automated” justice system, a bug is no longer just a “service ticket” but a violation of fundamental rights.
Key Takeaways:
- Understand the human biases we unconsciously transfer into systems.
- Analyze the documented failures of “judgmental” AI in the real world.
- Understand the danger of technological determinism and recognize the role of the “human-in-the-loop” through pop culture.
