AWS Certified AI PractitionerFoundation ModelsMedium
A developer is using a foundation model for code generation. When providing a prompt like 'Write a Python function to sort a list of integers,' the model consistently generates code that uses a bubble sort algorithm, even when more efficient algorithms (like quicksort or mergesort) would be better. This behavior is likely due to the model's pre-training data containing a disproportionately higher number of bubble sort examples. This scenario is an example of which ethical concern related to foundation models?
- ALack of explainability
- BIntellectual property infringement
- CAlgorithmic bias
- DHallucination
Show answer & explanationAnswer & explanation
Correct answer: C. Algorithmic bias
Algorithmic bias occurs when a model's outputs are systematically prejudiced or unfair due to biases present in its training data. In this case, the model's preference for bubble sort, even when less optimal, stems from an imbalanced representation of sorting algorithms in its pre-training data, reflecting a bias towards simpler or more common examples found online.
Why the other options are wrong
- A. Lack of explainability refers to difficulty understanding *why* a model made a decision, not the decision itself being biased.
- B. Intellectual property infringement would involve generating copyrighted code without attribution, which is a different ethical concern.
- D. Hallucination is generating factually incorrect but plausible content; here, the code is syntactically correct but suboptimal, reflecting a bias.
Algorithmic Bias (FM)
Systematic and unfair prejudice in the outputs of a foundation model, often stemming from biases present in its vast training data.
- Can lead to discriminatory or suboptimal outcomes.
- Manifests in various forms: gender, racial, cultural, or even technical preferences.
- Mitigation involves diverse data collection, bias detection, and ethical fine-tuning.
Memory trick: Bias Blocks Better Behavior.