Responsible Data Science

CHAPTER 9Conclusion

AI, and data science more broadly, bring the promise of seemingly unlimited good. After all, the ability to ingest any set of arbitrarily sized, minimally structured data and to produce predictions or explanations for these data is applicable to almost every domain. Our societal attention often focuses on the revolutionary future applications of this potential: cars that drive themselves, computers that can hold natural conversations with humans, precision medications tailored to our specific genomes, cameras that can instantly recognize any object, and software that can automatically generate new images or videos. However, we often forget that AI and AI-augmented methods already greatly empower us in the present. Authoritative information on almost any topic can now be accessed with the tap of a finger. Detailed photographs and video can be taken in near-complete darkness. Media and advertising platforms make startlingly accurate recommendations for the content or products that are most likely to appeal to us. Government services (ideally) become more cheap, effective, and fair. Whether we recognize it or not, these AI methods make an increasingly large number of decisions about our lives. These rapid, accurate decisions are often helpful: fraudulent credit card use is stopped in its tracks, home loan approval arrives quickly, and automatic content moderation removes illegal content before it is ever seen by a human. However, when these decisions are wrong, ...