kaal:claim:5095633-019

Biases held by human annotators or embedded in automated annotation systems are propagated into the models trained on their output, producing AI that performs inequitably across demographic groups.

there's a theoretical risk that biases inherent in data annotators or automated systems might be propagated into AI models. This can lead to AI that does not perform equitably across different demographic groups or scenarios.

Wulf A. Kaal, Artificial Intelligence The Final Frontier (2025), 4.1. Theoretical Shortcomings, p. 12
https://ssrn.com/abstract=5095633 · source PDF

Wulf A. Kaal, Artificial Intelligence The Final Frontier (2025). SSRN: https://ssrn.com/abstract=5095633

• applies even to providers that emphasize high-quality data such as Scale AI and Appen

mechanismsupport: arguedfailure: annotator bias propagationfamily: ai-model-and-training-failureinstitutional-design

The quote above is an exact substring of the source PDF, whose sha256 is cbb484711f89bcefc9fc6a5730a1ed0a3f764d7999ad9b6f7d8ea05634c26c63. Extraction method: pdf-text-layer.
Attestation record: colloquium/attestations/d3879add7b1e3b44...json
Verify the binding yourself: curl -s https://wulfkaal.github.io/claims/5095633-019.md | sha256sum