What is Algorithmic Bias? — ¿Qué es el sesgo algorítmico?

Francisco Martínez
5 min readJun 12, 2024

What is Algorithmic Bias?


Algorithmic bias refers to the systematically erroneous decisions that can arise in artificial intelligence (AI) and machine learning (ML) systems due to inherent biases in the training data, the design of the algorithm, or the way results are interpreted and used. This bias can lead to outcomes that are unfair or discriminatory towards certain groups of people.

Origins of Algorithmic Bias:

  1. Biased Training Data:
  • Description: If the data used to train an AI model contains biases or is unbalanced, the model will learn and replicate those biases. For example, if a historical hiring dataset reflects gender discrimination, a model trained on that data could perpetuate that discrimination.
  • Example: A facial recognition system that is predominantly trained with images of light-skinned individuals may perform poorly when identifying dark-skinned individuals.
  1. Algorithm Design:
  • Description: Decisions about how algorithms are designed, what variables are included, and how they are weighted can introduce biases. If an algorithm is not designed to mitigate biases present in the data, it can amplify them.
  • Example: A credit scoring algorithm that assigns excessive weight to certain socioeconomic factors may inadvertently discriminate against certain demographic groups.