Algorithm Bias: Computer Science Student Perceptions Survey

In the United States, Google performs over 3.9 million searches per minute. Monthly desktop searches can exceed over 10.7 billion and mobile searches are predicted to grow steadily. Concurrently, recent discourse has raised questions about bias in search engines and big data algorithms. As the information universe becomes increasingly dominated by algorithms, computer scientists and engineers have ethical obligations to create systems that do no harm. In this paper, the authors discuss a survey that was conducted of computer science and computer engineering students perceptions of algorithm bias. The aim of the survey was to gather preliminary data on how students perceive bias within machine learning and search algorithms. Over 700 computer science and computer engineering students from three different institutions participated in the survey from Fall 2018 to Spring 2019. Based on survey results, Google was overwhelmingly the preferred search engine. The participants also predicted that artificial intelligence algorithms will improve over time. The majority of respondents believe that private companies, not government organizations, need to regulate their own artificial intelligence algorithms. On average, computer science and computer engineering students acknowledge that algorithm bias could occur when people create algorithms. The results suggest that students are familiar with search engines and in general agreement on how algorithm bias should be addressed in the future.