EDon Conference

From Prof C (http://profcnews.com)

Transcript

1. Artificial Intelligence (AI) in The Academy and The World: ChatGPT J Scott Christianson
3. Arti fi cial Intelligence (AI) De fi nitions Arti fi cial Intelligence: A machine that exhibits cognitive or decision-making behavior and can take action to achieve a goal.
4. • General AI: A machine that can reason and adapt like a human. E.g, sci fi movies. Arti fi cial Intelligence (AI) De fi nitions Arti fi cial Intelligence: A machine that exhibits cognitive or decision-making behavior and can take action to achieve a goal.
5. • General AI: A machine that can reason and adapt like a human. E.g, sci fi movies. Arti fi cial Intelligence (AI) De fi nitions • : A machine that is optimized for a particular task or project. Narrow AI Arti fi cial Intelligence: A machine that exhibits cognitive or decision-making behavior and can take action to achieve a goal.
6. Machine Learning and Deep Learning Narrow AI Machine Learning Deep Learning Machine Learning’s goal is to develop predictions based on previously observed patterns. Various variables are weighed to predict the probabilities of the outcomes. The variables and formula used to make such predictions may be programmed by a human OR they can be developed by the machine itself (Deep Learning).
7. Deep Learning 01 Collect Training Data 02 Analyze and Segment 03 Setup and Train a Neural Network 04 Test and deploy
8. solid vertical diagonal horizontal e2eml.school Deep Learning Input Case 1 Case 2 Case 3 Case 4
9. solid vertical diagonal horizontal e2eml.school Deep Learning Input Case 1 Case 2 Case 3 Case 4
10. solid vertical diagonal horizontal e2eml.school Deep Learning Input Case 1 Case 2 Case 3 Case 4
13. Internet of Things (IoT) Artificial Intelligence (AI) A machine that exhibits cognitive or decision- making behavior and can take action to achieve a goal. Small, inexpensive, microprocessors, that collect environmental data and transmit that data to a central server for analysis or decision making !
14. cow4.jpeg cow3.jpeg
15. Types of Data For ML Processing • Motion Data • Audio/Voice Data • Image Data • Text Data • Geospatial Data • Physiological/Medical Data • Financial Data • Behavioral Data • and Much More
16. Types of Data For ML Processing • Motion Data • Audio/Voice Data • Image Data • Text Data • Geospatial Data • Physiological/Medical Data • Financial Data • Behavioral Data • and Much More
19. Figure 2: Applications of AI algorithms in medicine. The left panel shows the image fed into an algorithm. The right panel shows a region of potentially dangerous cells, as identi fi ed by an algorithm, that a physician should look at more closely. (From Arti fi cial Intelligence in Medicine: Applications, implications, and limitations by Daniel Green fi eld.)
20. • Radiology 396 • Cardiology 58 • Hematology 14 • Neurology 10 • Clinical chemistry 7 • Ophthalmic 7 • Gastroenterology and urology 5 • General and plastic surgery 5 • Pathology 4 • Microbiology 4 • Anesthesiology s4 • General Hospital 3 • Orthopedic 1 • Dental 1
21. Types of Data For ML Processing • Motion Data • Audio/Voice Data • Image Data • Text Data • Geospatial Data • Physiological/Medical Data • Financial Data • Behavioral Data • and Much More
26. Types of Data For ML Processing • Motion Data • Audio/Voice Data • Image Data • Text Data • Geospatial Data • Physiological/Medical Data • Financial Data • Behavioral Data • and Much More
27. Interactive Voice Response IF..Then Based Systems
28. Interactive Voice Response ML Based Systems
31. Digital Applications tend to improve at an exponential rate.
40. Types of Data For ML Processing • Motion Data • Audio/Voice Data • Image Data • Text Data • Geospatial Data • Physiological/Medical Data • Financial Data • Behavioral Data • and Much More
41. ~3 Years ago: AI Re-writers
44. ChatGPT is a poor student, but would pass.
48. Who Owns AI Generated Content?
51. Math is not a thinking tool used by most people. Writing is.
52. My Experience In my large lecture classes, I use short re fl ection paper assignments. Graded pass/fail (we might have 400 submissions). One of the options is to watch "Coded Bias" about the use of AI and facial recognition technology.
53. My Experience “Write a re fl ection paper about the fi lm coded bias” Coded Bias is a documentary fi lm that explores the ways in which arti fi cial intelligence and algorithms can perpetuate and amplify biases that exist in society. The fi lm delves into the history of AI and how it has been used in various fi elds, such as policing and hiring, and the consequences that have resulted from this use. One of the most striking aspects of the fi lm is the way it illustrates the widespread and often unconscious biases that are present in the data that is used to train AI systems. This includes biases against certain racial and ethnic groups, as well as biases based on gender and other characteristics. The fi lm shows how these biases can have serious consequences, such as the misidenti fi cation of innocent individuals as criminals or the unequal distribution of resources and opportunities. The fi lm also highlights the importance of accountability and transparency in the development and use of AI. It emphasizes the need for diverse representation in the fi elds of technology and AI, as well as the need for clear policies and regulations to ensure that AI is used ethically and fairly. Overall, Coded Bias is a thought-provoking and eye-opening fi lm that raises important questions about the role of AI in society and the potential consequences of its use. It serves as a powerful reminder of the need to be aware of and address biases in all aspects of our lives, including in the technology we use.
54. My Experience There are several problems with using training data to develop facial recognition systems: • Limited diversity: Many facial recognition systems are trained on data sets that are not representative of the entire population. This can lead to poorer performance for individuals who are not well represented in the data, such as people of color or those with disabilities. • Bias in the data: The data used to train facial recognition systems can contain biases that are inherent in the data collection process. For example, if the data is collected disproportionately from one racial or ethnic group, the resulting facial recognition system may be more accurate for that group and less accurate for others. • Data privacy: The use of facial recognition systems raises concerns about data privacy, as they rely on the collection and analysis of personal information. There is a risk that this information could be used for nefarious purposes or could be accessed by unauthorized parties. • Ethical considerations: The use of facial recognition technology raises ethical concerns, such as the potential for abuse of power or the erosion of privacy. It is important for the development and use of these systems to be guided by ethical principles and considerations “Explain the problems with training data and facial recognition”
58. Forbidden Fruit Some of us may have forgotten how hard it was to get a rough draft done when we were students.
59. Possible “Solutions” • Intense review of submitted assignments.
60. Possible “Solutions” • Intense review of submitted assignments. • Use of a detection tool (GPTZero).
62. —From NPR: https://www.npr.org/2023/01/09/1147549845/gptzero-ai-chatgpt-edward-tian-plagiarism
63. "The chatbot wrote medical research that managed to fool experts. In a recent study, medical researchers were asked to spot the fakes, and they identi fi ed 32% of the AI- generated abstracts as real and fl agged 14% of the genuine abstracts as phonies. Using computers to fi ght computers won't get us out of this mess—and AI output detector guessed wrong slightly more often than the researchers." —original article https://www.nature.com/articles/ d41586-023-00056-7
64. • Summary: "Rhetorical Situations and Their Constituents" by Keith Grant-Davie is a scholarly article that discusses the concept of rhetorical situations, which refers to the context in which communication takes place. The article explains that a rhetorical situation is made up of three main constituents: the rhetor (the person communicating), the audience, and the exigence (the problem or need that prompts communication). The article also examines how these constituents interact and how understanding them can help improve communication. Furthermore, Grant-Davie argues that understanding the rhetorical situation is crucial for e ff ective communication because it allows the rhetor to tailor their message to the speci fi c audience and exigence. The article also explores how the rhetor's purpose, the audience's expectations and attitudes, and the exigence's urgency and constraints all shape the rhetorical situation. • In addition, the author suggests that the audience and exigence can also in fl uence the rhetor's choices of discourse, such as their choice of words, tone, and style. The article also discusses how the rhetorical situation can change over time and how the rhetor must adapt their message accordingly. • Overall, "Rhetorical Situations and Their Constituents" provides a comprehensive overview of the concept of rhetorical situations and how understanding them can improve communication. The author emphasizes the importance of considering the audience and exigence in order to e ff ectively convey a message and achieve the desired outcome. • Connection: Im not sure i can connect in any way to this text other than wanting to further my understanding of rhetoric so i can better understand rhetorical situations when they occur. Result: Not AI Generated
65. • Summary: "Rhetorical Situations and Their Constituents" by Keith Grant-Davie is a scholarly article that discusses the concept of rhetorical situations, which refers to the context in which communication takes place. The article explains that a rhetorical situation is made up of three main constituents: the rhetor (the person communicating), the audience, and the exigence (the problem or need that prompts communication). The article also examines how these constituents interact and how understanding them can help improve communication. Furthermore, Grant-Davie argues that understanding the rhetorical situation is crucial for e ff ective communication because it allows the rhetor to tailor their message to the speci fi c audience and exigence. The article also explores how the rhetor's purpose, the audience's expectations and attitudes, and the exigence's urgency and constraints all shape the rhetorical situation. • In addition, the author suggests that the audience and exigence can also in fl uence the rhetor's choices of discourse, such as their choice of words, tone, and style. The article also discusses how the rhetorical situation can change over time and how the rhetor must adapt their message accordingly. • Overall, "Rhetorical Situations and Their Constituents" provides a comprehensive overview of the concept of rhetorical situations and how understanding them can improve communication. The author emphasizes the importance of considering the audience and exigence in order to e ff ectively convey a message and achieve the desired outcome. • Connection: Im not sure i can connect in any way to this text other than wanting to further my understanding of rhetoric so i can better understand rhetorical situations when they occur. Result: AI Generated
66. Possible “Solutions” • Intense review of submitted assignments. • Use of a detection tool (GPTZero). • Problems: These tool are not validated, created by individuals, and don’t provide sources/references to back up probable use of GPTChat. Also post-generation modi fi cations to text can provide bad results.
68. Possible “Solutions” • Intense review of submitted assignments. • Use of a detection tool (GPTZero). • Problems: These tool are not validated, created by individuals, and don’t provide sources/references to back up probable use of GPTChat. Also post-generation modi fi cations to text can provide bad results. • Use Google Docs with Tracking Changes, so you can see the student’s work evolve. Large scale pasting of text can be seen. “But, I used Pages and then pasted into Google.” • Have students write the text in class. I’m back baby !
69. Drawing the Line • Grammarly is AI Driven and used by faculty and students.
70. Drawing the Line • Grammarly is AI Driven and used by faculty and students. • MSWord/outlook does sentence completion.
71. Drawing the Line • Grammarly is AI Driven and used by faculty and students. • MSWord/outlook does sentence completion. • What about Lex.page ?
72. Drawing the Line • Grammarly is AI Driven and used by faculty and students. • MSWord/outlook does sentence completion. • What about Lex.page ?
73. Drawing the Line • Grammarly is AI Driven and used by faculty and students. • MSWord/outlook does sentence completion. • What about Lex.page ? • Every computer writing tool will use AI to some degree • There are more tools everyday: https://theresanaiforthat.com • Faculty are using ChatGPT for writing grants, professional development plans., etc.
75. A priori A posteriori
76. Possible “Opportunities” • Don’t worry, be happy. “Mollick’s new policy states that using A.I. is an "emerging skill"; that it can be wrong and students should check its results against other sources; and that they will be responsible for any errors or omissions provided by the tool.”
 —NPR • Use ChatGPT/AI to draft outlines • Use ChatGPT/AI to summarize material • Use ChatGPT/AI as a tutor
77. Possible “Opportunities”
78. Possible “Opportunities”
79. Agenda •Why are we here? •What is GPT/ ChatGPT? •Student motivations. •Possible “solutions.” •Where do we draw the line? •What’s Next? •Trending or Trendy? •Q&A
80. Forget ChatGPT. Worry about AI. • AI is already used for grading. • AI is already used at some institutions for screening admissions. • AI based facial recognition used on some campuses. or Trending Trendy?
81. When and How to Use ML • Autonomous Vehicles AI and Ethics
88. When and How to Use ML • Autonomous Vehicles • Admissions and Grading AI and Ethics
89. When and How to Use ML • Autonomous Vehicles • Admissions and Grading • Loans and Credit AI and Ethics
90. When and How to Use ML • Autonomous Vehicles • Admissions and Grading • Loans and Credit • Social Media AI and Ethics
91. When and How to Use ML • Autonomous Vehicles • Admissions and Grading • Loans and Credit • Social Media • Warfare AI and Ethics
94. Problems with AI
95. solid vertic diagonal horizontal e2eml.school Input Case 1 Case 2 Case 3 Case 4 Problems with AI “Hidden Layers”
96. Problems with AI Adversarial AI from Savan Visalpara
97. Problems with AI Adversarial AI from Weijia Zhang
98. Problems with AI Adversarial AI from MIT CSAIL
99. Problems with AI Adversarial AI Images: Evtimov et al Camou fl age graf fi ti and art stickers cause a neural network to misclassify stop signs as speed limit 45 signs or yield signs.
100. Volume Velocity Variety Veracity
101. Questions and Answers ProfCnews.com