Download OIDD X: AI, Business, and Society ( CU) and more Study notes Artificial Intelligence in PDF only on Docsity!
OIDD 255X: AI, Business, and Society (1 CU)
Instructor: Prasanna (Sonny) Tambe
Last updated 10/22/
SAMPLE SYLLABUS, ALL ELEMENTS ARE SUBJECT TO CHANGE.
Course Objectives
The progression of AI-based technologies promises to transform many aspects of business, labor, and even society. The goal of this course is to provide students with an understanding of the capabilities of modern AI technologies, with an emphasis on being able to critically assess where they can provide societal value, and where they may create new societal challenges. The course is not intended to provide a deep-dive into the workings of these technologies in the same way as a computer science course might. Rather, business and policy decision-makers will be confronted with a number of important issues as AI becomes integrated into the social decision-making fabric. This course is intended to provide a framework for people who may have to confront these legal, ethical, and economic challenges. In doing so, a course objective is to ensure that students who complete the course are comfortable enough in the inner-workings of these technologies to think critically across many AI contexts as well as different domains ranging from public policy, to criminal justice, to health inspections, HR, and marketing.
This 1 CU course is oriented around hands-on critical written assessments, labs, exams, and a presentation. Broadly, data rich firms in finance, tech, management, marketing, and other industries are increasingly adopting AI as a tool to accelerate and improve decision-making. It is important for modern managers to understand the opportunities and challenges introduced by data and AI so that they can credibly communicate about these issues with others in the firm. We will cover many of these issues, so that you will be able to think about the opportunities and challenges that arise when firms try to use AI to solve business problems.
Labs will reinforce your learning of how AI works, and how it is being used to solve business problems. We will focus in the labs on gaining experience with introductory machine learning concepts. Students will spend time inside and outside of the classroom combining data and code to provide a foundation for understanding the deep challenges that this will bring to organizations.
Course Overview
Within the last decade, there has been a dramatic rise in interest in the use of AI technologies applied to many domains, including finance, HR, policy, marketing, and strategy. As a result, the modern “digital leader” increasingly requires the use of technology, statistics, and data analysis skills to facilitate business analysis as well as a grasp of where AI technologies can perform well, and where they may fail. This includes knowing how to a) effectively frame data-driven questions, b) use AI-driven algorithms, and c) understand the capabilities of this new generation of tools that are becoming available to automate decision-making from data.
The class includes readings and critical assessments due before class and in-class discussions of topics related to the application of AI technologies, including ethics, bias, and the potential for AI to fuel gains in productivity. The class also requires the completion of two labs to ensure that
participants have a deep understanding of how algorithms are applied for decision-making and what the constraints are of these approach. A background in coding is not required in this class. Labs will be based in “no-code” tools such as MS Azure ML Studio.
Throughout the semester, we will cover the applications of AI to a number of industries (e.g. medicine, journalism, criminal justice), including inviting guest speakers to hear about applications in special- ized domains. A learning goal of this course is exposure to how AI is changing decision-making in different industry contexts, and how organizations are reacting to these changes.
Class time in most weeks will be dedicated to lecture and discussion of some of the most important topics facing the AI community. Class time in some weeks will be devoted to supervised work on projects that are meant to underscore AI-based challenges. Through these exercises and discussions, students are expected to become proficient at applying data to business decisions and at effectively analyzing big data sets to inform decisions about business problems using data analysis tools.
Course web site
We will be using Canvas to submit all assignments and receive grades. All course information will be posted on the Canvas website.
Required textbooks and software
There is no required textbook. There are frequent readings that will consist of selected online content which will be posted on the course site. As part of your homework, you may be expected to download and install some open source software.
Deliverables and grading
During this course, you will be assigned a number of hands on data projects which you will spend time on both in class and out of class. You are expected to participate in classroom discussions (there is more information below about participation). The breakdown of points is as follows:
Deliverable Weight Points AI presentation 5% 25 Labs 5% each (x 3) 75 Response writeup 5% each (x 2) 50 Midterm 15% 75 Participation 15% 75 Final Exam 25% 125 AI pitch deck 15% 75 TOTAL 100% 500
With each project, you will be provided with a set of guidelines. Deliverables may include short, informal analyses and an accompanying recommendation. Group projects will be completed in small groups (two to three students, no more than three). You may also be asked to evaluate the contribution of each of your team members after the group project.
Schedule
Session Date Topic Due 1 xxx Course Intro 2 xxx AI: The past to the present 3 xxx AI Applications: Presentations Presentations 4 xxx AI Applications: Presentations 5 xxx Understanding Digital Data 6 xxx How AI works: ML 7 xxx How AI works: Loss functions 8 xxx In class ML Lab 9 xxx How AI works: Deep learning ML Lab 10 xxx In class Deep Learning Lab 11 xxx Quiz Quiz 12 xxx Ghost Work: How Data Gets Labeled 13 xxx Reinforcement Learning 14 xxx Generative AI 15 xxx AI Competition: Software, Skills, Data, and Computation DL Lab 16 xxx AI Competition: Networks and Regulation 17 xxx AI Barriers: Bias Auto ML Lab 18 xxx AI Barriers: Explainability 19 xxx Industry analysis: HR + Guest 20 xxx AI Barriers: Ethics & Law Bias response 21 xxx AI Barriers: Privacy 22 xxx Industry analysis: Healthcare + Guest 23 xxx AI Culture and Governance Ethics response 24 xxx AI & Jobs I 25 xxx AI & Jobs II: Scenario exercise 26 xxx Exam Final exam 27 xxx Final project presentations 1 28 xxx Final project presentations 2 Final project
Individual Session Descriptions
Session 1: Course Introduction
Session Overview : Introduction to course essentials, including topics and concepts covered, session structure, and grading. Group-led discussion of critical issues in AI.
Session 2: AI: The past to the present
Session Overview : Surveys a “history” of artificial intelligence, ranging from the 1800’s to modern day. Discusses the Turing Test, the AI winter and spring, expert systems, and the rise of modern statistical machine learning.
Session 3: Presentations - AI applications
Session Overview : Groups make short presentations on modern applications on AI issues.
Session 4: Presentations - AI applications
Session Overview : Groups make short presentations on modern applications on AI issues.
Session 5: Understanding digital data
Session Overview : Covers the foundational technical knowledge required for understanding how raw digitized data is connected to progress in AI. Includes bits and the digitization of information, and the role of tranistors and Moore’s Law.
Session 6: How AI works - Machine Learning
Session Overview : An introduction to ML techniques, with comparisons with pattern recognition and regression. Introduction to how prediction works in the ML context. This session covers how to think about different popular ML algorithms (logistic regression, trees, etc.).
Session 7: How AI works - Loss functions
Session Overview : Discussion of how to evaluate AI models and further discussion of the costs and benefits of different AI models. Explanation of ROC curves, false positives and false negatives, and confusion matrices. Discussion of how the need for well labeled output limits which tasks are amenable to machine learning/automation. Will also cover platforms such as Kaggle, and how ML competitions are “specified”, including prediction quality measures. Finally, we will walk through an example of how no-code tools are used for machine learning.
Session 8: In class Machine Learning lab
Session Overview : Use of the open source WEKA data mining tool to complete an empirical case study on the use of machine learning to predict HR attrition in the IBM employee flight data set.
Session 9: How AI Works: Deep Learning
Session Overview : Introduction to deep learning and how it works. Contrasts the benefits of hand- generating ML features vs. using a deep learning engine. Covers feed forward networks and back propogation.
Session 17: AI Barriers: Bias
Session Overview : Covers bias in ML algorithms. Included discussions of tradeoffs between ac- curacy and bias and questions around high profile cases such as the OPTUM system and the ProPublica/Northpoint application to criminal justice.
Session 18: AI Barriers: Explainability
Session Overview : Discusses problems related to explainable AI and why explainable AI has become a major focus of AI development. What are the trade offs in organizational adoption? What new technologies are emerging to make AI decisions more explainable?
Session 19: Industry Analysis - HR
Session Overview : Walks through a deep case analysis of an application of AI to HR applications, with an emphasis on challenges related to deployment, bias, interpretation, and legal issues that arise when applying AI to HR. Likely to involve a guest visit to provide industry context.
Session 20: AI, Ethics, & the Law
Session Overview : Discusses the challenging issues around AI ethics. This includes the question of AI morality and integrates examples such as the Trolley problem. Half of the class session is dedicated towards an in-class debate.
Session 21: AI & Privacy
Session Overview : Covers challenges at the intersection of AI and privacy, including the privacy paradox, legislating whether consumers must be required to “opt-in”, relevant case law, and differ- ences across countries. Also covers differences in state-led and market-led approaches to managing information privacy as well as implications for competition in the technology industry.
Session 22: Industry Analysis - Healthcare
Session Overview : Walks through a deep case analysis of an application of AI to healthcare applica- tions, with an emphasis on challenges related to deployment, bias, interpretation, and legal issues that arise when applying AI to healthcare. Likely to involve a guest visit to provide industry context.
Session 23: AI Management & Governance
Session Overview : Covers how organizations must change to integrate AI-supported decisions. Includes issues around governance (e.g. the role of AI councils), challenges with AI-led management, and how management may be different in organizations that rely heavily on data-driven decision- making. Discusses data-driven culture.
Session 24: AI & Jobs I
Session Overview : Discussion of the impact of AI on the future or work. Includes job displacement, robots, universal basic income, and other related topics. Discusses how the workforce and organi- zations are changing to adapt to AI. Also discusses what types of managers are needed to govern AI driven decisions, including the emergence of new governance vehicles, such as AI councils and Chief Data Officers.
Session 25: AI & Jobs II
Session Overview : Conducts a scenario planning exercise. Students are divided into groups and asked to “game out” scenarios and policy responses under different assumptions about the rate and direction of AI-led job displacement. In the last twenty minutes, students present back to the class their policy conclusions under the different scenarios.
Session 26: Final Exam
Session Overview : Final exam covering all concepts covered through the course, with an emphasis on those covered in the second half.
Session 27: Final Project Presentations I
Session Overview : Students present their “pitch deck” for a new AI driven venture.
Session 28: Final Project Presentations II
Session Overview : Students present their “pitch deck” for a new AI driven venture.
