Online products that a user can interact with for purposes other than retrieving information.
Author: University of Edinburgh; Publisher: edX; Publication Year: N/A. The following online class was created by the University of Edinburgh which is available for free on edX. The course is 7-weeks long and requires only 3-4 hours of work per week, but the course covers a large variety of material relating to data ethics, artificial intelligence (AI), and responsible innovation. The course focuses on educating the student on…
Author: Timnit Gebru; Publisher: DAIR Research Institute; Publication Year: N/A. The following website features an interdisciplinary and globally distributed artificial intelligence (AI) research institute that was founded by Timnit Gebru who was fired by Google in December 2020 for raising issues of discrimination in the workplace as was was the co-lead for Google’s Ethical AI research team. This resource seemed unique because it’s…
Author: N/A; Publisher: The Data Nutrition Project; Publication Year: N/A. The following organization’s mission is to empower “data scientists and policymakers with practical tools to improve [artificial intelligence (AI)] outcomes.” This page outlines the problem with AI algorithms as garbage in, garbage out. They argue that training datasets need to be assessed based on standard quality measures that are both qualitative…
Author: Miro Dudik; Publisher: Fairlearn; Publication Year: N/A. The following tool is Fairlearn which is a open-source project that helps data scientists improve fairness of their artificial intelligence (AI) systems. It includes a python library to assess fairness and improvement. It also includes educational resources about technical processes for unfairness mitigation. Key components of the fairness package are a…
Author: Mehrnoosh Sameki; Publisher: Microsoft; Publication Year: 2021. The following dashboard can be used to make sure data scientists ethically and responsibly take the correct approach towards artificial intelligence or machine learning. “A single pane of glass bringing together several mature responsible AI tools in the areas of machine learning interpretability, unfairness assessment and mitigation, error analysis, causal…
Author: Maria Tackett; Publisher: Data Science Box; Publication Year: 2019. The following slideshow is a good resource to go through for data professionals because it’s easy to follow and makes a great point. Even well intentioned people or visualizations can be unethical in how the data is presented and understood by the normal viewer….
Author: Mando Rotman, Tom Jongen, Floor Komen; Publisher: IG&H; Publication Year: N/A. The following tool is an Ethical Risk Quick Scan for helping clients assess the ethical risks of their data. The following questions are asked during the Quick Scan: 1). Vulnerable people impacted; 2). Number of people impacted large; 3). Matters of life affected; 4). Influencing personal behavior; 5). No, or slow, feedback loop; 6). No human in…
Author: Joy Buolamwini; Publisher: The Algorithmic Justice League; Publication Year: N/A. The following group is an organization that addresses the social implications and harms of artificial intelligence (AI) by raising public awareness about the impacts of AI and work with advocates and communities to mitigate AI bias and harms. They look at equitable and accountable AI, and are focused on human rights and harms caused by AI. Their values…
Author: Iyad Rahwan, Jean-Francois Bonnefon, Azim Shariff; Publisher: Moral Machine; Publication Year: N/A. The following resource is an interactive tool that presents the user with several “trolley problem” type situations in which an autonomous vehicle will crash into 1 of 2 groups of people, varying in age, gender, profession, etc. The tool aims not only to crowd-source public opinion on the specific situations presented but to start a conversation about…
Author: H.V. Jagadish; Publisher: University of Michigan, Coursera; Publication Year: N/A. The following course from the University of Michigan on Coursera. It is primarily an introductory course that provides the history of the concept of informed consent, data ownership, privacy, anonymity, and data validity. The course also offers brief exposure to the notion of algorithmic fairness that is sometimes violated given the unfair assumptions of an…