The following article summarizes the original event which is listed below the summary.
What is artificial intelligence?
What do hallucinations, articles on rectangular land surveys, and pirate ships in coffee cups have in common? They all can be generated by artificial intelligence (AI). UMRA members learned that and much more about generative AI from our April 23 Forum guest speaker, University of Minnesota School of Law Professor Daniel Schwarcz.
Generative AI is a form of machine learning. It is different from the computer programs we are familiar with that involve setting rules for computers to follow. Instead, with generative AI, humans feed extremely large amounts of data into a virtual neural network and the program teaches itself in ways that are not totally understood.
The goal, when creating text, is to generate a series of words that make sense. A training algorithm provides feedback to the virtual neural network and, eventually, a coherent sentence and paragraphs are created. Human feedback enhances the learning.
The end result is a language prediction model that remains static but is used to generate responses to questions. Currently, the Generative Pre-trained Transformer or GPT-3.5 and GPT-4 versions are in use. GPT-4 was launched in March 2023 and is 10 times more advanced. A new, greatly improved version is likely to be released soon, Schwarcz said.
AI is best at creating summaries of existing material, Schwarcz said. He demonstrated this by asking the application for information on rectangular land surveys in the Minnesota territory, as suggested by UMRA member Rod Squires. Squires, a scholar of land surveys, was less than impressed with the depth of the resulting summary but agreed it was accurate.
Hallucinations
Schwarcz said that when AI can’t find information it creates incorrect predictions called “hallucinations.”
Schwarcz has conducted research on student use of AI. In his work, students who were at the low end of the grade distribution benefited from use of AI, while those at the upper end were hurt. He thinks that is because AI replaces independent thinking and requires the student to take time to edit and remove hallucinations. Similar research involving writing classes showed AI did not improve the quality of the work product, but students were 20-30 percent faster.
Other research presented various standardized exams (AP, GRE, LSAT, and bar exams) to AI, and compared the scores with the results from actual people. Two years ago, AI generally scored better than 20-30 percent of the actual test takers. When the research was repeated one year ago, AI often scored better than 90 percent of the people. This shows the speed at which AI is improving, Schwarcz said.
Generative AI can also create non-text output and, for a bit of humor, Schwarcz showed several animated images created by AI based on his prompts. The video of pirate ships fighting in a coffee cup is one that will stick in my mind.
So, is generative AI a good thing or a bad thing?
Schwarcz did not provide a definitive answer, but he did raise numerous questions about AI’s future.
- Will its exponential growth rate in learning continue?
- Do we have enough data to teach AI?
- How does copyright law protect people whose data (writing) is being fed to AI models?
- What will AI do to educational and cultural attitudes about expertise and hard work?
- Will employers use AI to support or replace humans?
- How do we keep humans in the loop when AI makes critical decisions?
- How do we regulate AI to prevent harm?
The answers will determine whether we rue the day AI was let out into the world or praise the technology for its incalculable benefits.
—Julie Sweitzer, UMRA president-elect and Program Committee chair
What is generative AI and is it a good thing or a bad thing?
Tue, April 23 2024, 11am
Daniel Schwarcz, JD
Fredrikson & Byron Professor of Law
University of Minnesota Law School
Midland Hills Country Club
2001 Fulham Street
Roseville MN 55113
Artificial Intelligence (AI), in general, consists of programming a computer to do tasks. “Generative” AI is the idea that a computer can scan a database containing many items, such as words, and then create new content by predictively combining selections of those words into an answer to a prompt.
A very simple, first-generation illustration is the text management system on your smart phone. Have you noticed that when you are typing a text message, your phone will suggest words to you to complete your thoughts? This is the computer predicting the content of your current message based on a scan of your word use in similar messages previously. Now, radically expand that concept to a computer scan of massive databases of words across the Internet, at the end of which the computer will create an answer to a question you ask.
Substantial debate
Generative AI has attracted substantial debate in that it can be used in many contexts in many fields, including in educational settings to do things such as answer exams and write papers. Imagine, for example, that a student enters a prompt into a computer, the relevant AI programming tool collates and synthesizes a vast stockpile of data into a narrative form, and out comes a work product. Is this good? Is this bad? And how do we know?
One can argue that generative AI promotes efficiencies in the creation of documents. One can also counter argue that its use crowds out demonstration of independent knowledge and judgment on the part of the researcher/writer. And how do we know whether, in its assembly process, the computer has created accurate content?
In this forum at 11 a.m. on Tuesday, April 23, Professor Schwarcz will talk about what generative AI can and cannot do, discuss effective prompting strategies, identify methods for verifying the accuracy and reliability of AI-generated output, and address the question of “dumbing down” the quality of work through over-reliance. He will also talk about how tools such as ChatGPT—released in 2022 and credited with sparking the AI boom—work in comparison to other versions of AI.
Schwarcz is the Fredrikson & Byron Professor of Law at the U of M Law School. He studies a broad range of issues in the domain of insurance law and regulation, including the use of AI by insurers. He is an award-winning teacher and scholar. His work has received the American Law Institute’s prestigious Early Career Scholars Medal. He is regularly covered in media outlets including The New York Times, the Wall Street Journal, and National Public Radio, and he has testified to U.S. congressional committees. Professor Schwarcz earned his AB, magna cum laude, from Amherst College and his JD, magna cum laude, from Harvard Law School.
Do not miss this forum! Make your reservation for UMRA’s April 23 luncheon forum at Midland Hills Country Club in Roseville today.
—Brad Clary, UMRA Program Committee
Upcoming Events
We are fortunate to have UMRA member and former Minnesota Landscape Arboretum director Peter Moe to lead this approximately 3.7-mile hike at the Arboretum on Monday, May 6. Plan to have lunch together after the hike at the Eatery café inside the Oswald Visitor Center.
This session will continue the discussion begun in the May 8, 2023 session. See that article for details.
If you currently live in your own home or apartment, have you looked ahead to when you might want to consider a senior living option? Do you have any idea what those options might be? How much they cost? What they provide? UMRA’s May 14 workshop will provide an overview of the world of senior living and the basic information you need.
You’re invited to join the final Family History Interest Group meeting for 2023–24. This will be an opportunity to ask questions related to family history, share insights, talk about recent projects, and suggest topics for next year.
Stephanie Daily will lead the discussion of Fuzz: When Nature Breaks the Law by best-selling author Mary Roach when the UMRA Book Club meets via Zoom at 2 p.m. on Friday, May 17.
The May 20 hike will start at the Historic Lift Bridge in Stillwater and go along the river to the new bridge, cross over and then back on the Wisconsin side, a total of approximately 5 miles. Most of the route is paved. We will eat lunch after at Brian's Bar and Grill in Stillwater. Car pooling will be available.
Climatologist and meteorologist Mark Seeley joined the U of M faculty in 1978 and has been a weekly commentator on Minnesota Public Radio since 1992. We are honored to welcome him as our guest speaker for UMRA’s May Luncheon Forum and 2024 Annual Meeting. He is always entertaining and informative.
UMRA’s Fourth Friday Book Club welcomes back, for the second time, author Julie Schumacher, U of M Regents professor of creative writing and English, for a conversation about her novel The Shakespeare Requirement.
We are in for a treat on June 3! Lynn Anderson, will lead us on trails at two parks near her Monticello home. Bring a bag lunch and we will gather in Lynn's yard along the Mississippi River for eating and socializing after the hike.
Get out to the ballpark with fellow UMRA members for the annual U of M Day at Target Field as the Minnesota Twins take on the Oakland Athletics on June 14 at 7:10 p.m.
Please join us at 9:30 on June 17, 2024 for a walk around Lake Como in St. Paul before attending the UMRA Summer Social and Picnic starting at 11 at the Como Pavilion. A perfect opportunity to get in a walk before we gather to celebrate summer.
Picnic under the covered, open-air Como pavilion, gaze at the water, spy some birds, catch up with friends and former colleagues, and meet new members. We’ll have lunch with plenty of time for visiting and then compete in teams as we are challenged by our Quiz master(s)!
Healthy aging, mindfulness, fighting ageism, and more will be the focus of the third annual Age-Friendly University Day to be held on the U of M Twin Cities campus.