In Data Science, algorithmic bias refers to errors in the results of machine learning algorithms due to erroneous assumptions in the implementation of the solution. Several factors can influence algorithmic bias, some of them refer to the data presented, whereas others are a result of bias that humans already have. It’s important that we start this conversation around bias in data science to allow us to get ahead of the game and prevent as much of it as possible. Machine learning algorithms learn from data, so, if data is biased then there is a high chance that the algorithms from which it came are too. Furthermore, algorithms are created by humans, and if those humans composing the data science team are biased, then the algorithm most likely will be too. 

So, the more diverse the data science teams are the more of a chance we have to prevent bias in emerging technologies. 

Takeaways:

• Bias is unavoidable 
• Bias is present in the data 
• Bias is present in people