AI: Facing the Fear

Fear of the unknown. It has been with humankind since before a cave dweller first struck fire. But just as that first fire brought warmth, security and lit up the landscape it also brought the potential for injury and destruction.

That conundrum continues today with the advent of Artificial Intelligence. It has the potential to throw light into our lives but it can also consume us. “A lot of people are frightened about it,” said Gretchen Hachmeister, executive director of the Hotchkiss Library in Sharon.

The library will undertake the task of enlightening the public about AI this Saturday at 11 AM when it offers, Understanding AI: What It Is and What It's Not, the first of two programs on how to navigate the rapidly evolving world of artificial intelligence.

It will be a beginner-friendly program that will separate fact from fiction about AI technology and will be followed on Saturday, September 6th, by AI: Prompting and Practical Tools, Chat GPT, Perplexity and Others.

The series will be presentation-style with visuals and real-world examples, and is ideal for adults with no experience with AI.

“AI can’t be avoided so the only way to not fear it is to understand it,” said Hachmeister. “We are confident these programs will help demystify what it can and cannot do. The second part will be on how to use it and be sure it’s fairly reliable.”

The programs are an outgrowth of the community’s own desire. “We did a community survey this spring about the library and its offerings,” said Hachmeister. “One of the top requests was for technical training and a number of people mentioned AI. There is a great deal of information bombarding us here at the library but I didn’t know much about it, so I started reading.”

Hachmeister said people see libraries as a point of contact about technology. “There are a lot of things we can do to help people but others we cannot,” she said. “We are in the early stages of exploring what we need to know about AI and are starting professional development on the staff side.”

She said the public reaction to the advent of AI has been mixed. “I had one person tell me that if we use AI he will never come into the library again. Others are curious about it and want to know more. People are frightened of it but it makes fact checking much easier.”

The programs will explore how generative AI tools like ChatGPT function, what their limitations are and how they are trained. They will also address copyright, privacy, misinformation and ethical concerns including algorithmic bias.

The presenter will be Pam Doran, a SUNY Empire 2021-22 PRODiG Scholar and doctoral student. She is an ISTE-certified educator and trainer, a certification that recognizes proficiency in implementing technology to enhance teaching and learning outcomes.

She is also a Certified Professional in Accessibility Core Competencies (CPACC), a credential that demonstrates her comprehensive understanding of digital accessibility principles, laws and best practices and serves as the Digital Accessibility Coordinator at SUNY Empire State University. Currently, she is working with Dr. Heather Horowitz and Dr. Jeff Foulkes on a project titled Enhancing Student AI Literacy through Universal Design for Learning.

While she is well-versed in the emerging technology—indeed, is developing some of its programs—she understands the “dark underbelly” of relying on general AI programs. In an online class about the capabilities of Chat GPT, Doran described it as a “dicey subject.”

ChatGPT is built to engage in natural language conversations, meaning it can understand questions and respond in a way that feels like a human interaction. It can be helpful in doing research, Doran said, but it must be put in a “supporting role.”

“AI is an assistant and not a replacement,” she said. “It is a useful resource to generate tools and to locate other sources but it cannot replace my critical thinking, my domain expertise or learning—and I most definitely don’t like it replacing my writing. The idea of a robot writing for me is repugnant—I’m kind of indignant about that.”

Because ChatGPT is conversational, she says she uses it to debate concepts she is working on. She will assign it a side in a debate and bounce ideas off it. “That’s when you go deeper, you go through a Socratic conversation with it. Having a dialogue with an opposite voice is helpful in your clarifying voice. That is one of the things I do with ChatGPT.”

She issued caveats about using AI, however, noting that it has a “tremendous potential” for interjecting bias into the information it presents. “It’s gathering information from the Internet, which is biased,” she said.

She further cited accuracy concerns, saying AI may produce “inaccurate or fabricated information,” sometimes known as “hallucinations.”

“You need to validate your sources and employ critical thinking—you treat the references as prompts but you have to question their authenticity,” she said. “It can blur the lines of original authorship and we need to be concerned about data privacy.”

“We hope this series will offer information about which applications check their sources,” said Hachmeister. “It’s important to use applications that tell you where they get their information. Other applications will tell you whether you can use the information legally. There is a hearing in the Senate (this week) about copyright protection.”

Lastly AI, because it is generated by emotionless robots, is subject to “context blindness.” “It does not understand context the way humans do,” Doran said.

Those interested in attending should register for each of these sessions at the library’s website hotchkisslibraryofsharon.org. For more information call the library at 860-364-5041.