I am sure you have heard the term “you are what you eat”, so I hope you have been eating lots of fruits and vegetables (on a side note, the human body is 65% water, so drink water). Just like we need quality food to be healthy, data science models need quality information to be useful and serve everyone that has a stake in the project. As a data scientist, we are responsible for finding this information and deciding what is useful and what is not. For example, remember Tay? If you don’t, Tay was an AI bot created by Microsoft that was supposed to mimic the lingo of a teenage girl on twitter. Tay was supposed to learn from the interactions it would have through twitter. In less than 24 hours, Tay had become racist. Microsoft quickly shutdown Tay and tweaked it before it relaunched it. Besides embarrassment and damaging the brand, the AI did not do a lot of real-world damage. However, there are court systems that use AI to arrive at a jail sentence. Unfortunately, this AI disproportionately gives minorities harsher jail sentences. This AI does great damage to the community because it continues to enforce racist policy. I am sure that the results weren’t intended, but it is still something that should have been brought up at some point. Both AIs have something in common, they were fed bad information. I am sure that the development team did not intend for either of those cases to happen, but it’s something that should have been accounted for. How can we possibly account for everything? We can’t, but we can take a step closer by having diverse development teams. We need more women, more minorities, and more people from different groups to help shape the development of AI. This is the only way that we can ensure that the future of AI serves everyone fairly.