AI Ethics

Catherine Gracey, Open Scholarship and Applied Sciences Librarian

ENGG-1002, April 2025

Hello!

• I'm the Computer Science (and Entrepreneurship) librarian, and I work on a lot of the library's AI projects
• I also research student use of AI, and AI search tools

Today's Talk

1. Searching
   • Chatbot VS Retrieval Augmented Generation (RAG)
   • What corpus/collection is the tool searching?
2. Writing 
3. Coding
4. Art/Media
5. Transcription

Artificial Intelligence for Undergraduate Students Guide

Different tools have different use cases

While this is BY FAR the most popular tool with students, it has some key issues, and there are alternatives available to UNB students.

1. Lack of references/sources, combined with a definitive tone
2. Academic sources are presented on equal level as pop sources, like blogs, magazines, etc.
3. Huge difference in quality between free/paid versions

I used a research question from a UNB PhD student's dissertation and tested it in ChatGPT and ScopusAI

ChatGPT

(free version)

ScopusAI

(via UNB)

ChatGPT

(free version)

ScopusAI

(via UNB)

ChatGPT

(free version)

ScopusAI

(via UNB)

• GenAI is a product of its training data
• It WILL adopt the biases of the data/people who created it
• This means it can reflect real-world prejudices and biases
• Without constraints, this can harm people

[1]

[2]

[3]

[4]

[5]

(nothing is - let's talk tech)

Netflix Pricing

Building a loyal customer base

We are here!

If you are not paying for it, you’re not the customer; you’re the product being sold.

– Tim O’Reilly / Andrew Lewis / Richard Serra and Carlota Fay

• Not in use in most AI tools now, but a likely future revenue stream
• Advertisers are the customer, not you, they are paying tools for your time
• Think Tiktok, Instagram

Advertising

• The 'content' you input to AI can be used to train models (almost always in free models)
• OR, your data could be sold to advertisers, like in the Cambridge Analytica Facebook Scandal

(If we let it)

• Don't freak out! We still need engineers!
• But! You need to maintain key skills and cognitive functions, while also learning to use AI

Plagiarism includes:

(a) quoting verbatim or almost verbatim from a source (such as copyrighted material, notes, letters, business entries, computer materials, etc.) without acknowledgment;

(b) adopting someone else’s line of thought, argument, arrangement, or supporting evidence (such as, for example, statistics, bibliographies, etc.) without indicating such dependence;

(c) submitting someone else’s work, in whatever form (film, workbook, artwork, computer materials, etc.) without acknowledgment;

(d) knowingly representing as one’s own work any idea of another;

(e) contravention of written instructions of the instructor dealing with plagiarism.

Academic Integrity and Artificial Intelligence (AI2)

1. Refer to the specific policy for your class/professor
   • This means the rules can change from class to class
   • This can be confusing, so don't be afraid to ask
   • Professors are also doing their best, and trying to help you develop the skills they think are essential 
2. Never misrepresent AI-generated work as your own
   • See Academic Integrity Guide

What this means practically:

So which skills matter? What can you outsource to AI?

These are questions you should be asking yourself the rest of your time at UNB. A resource to get you started:

[1]

[2]

Professionals who utilize AI are ultimately responsible for its impacts

• It can't be held legally responsible for mistakes in the same way humans can
• This is a really complicated area! As tech/AI becomes more integrated this might change
• But for now, you must use due dilligence to ensure key mistakes aren't made (think Iron Ring!)

AI does not have legal personhood

• Remember that AI companies are businesses, and that you have power as a consumer!
• Diversify your AI tool use (go beyond ChatGPT)
• Verify information from AI outputs
• Cite sources directly, not just the AI tools, this is like saying "I got this from books" or "Google"
• Research what professional engineers are saying about AI, and the skills you'll need
• Keep building skills away from AI tools (make sure you can do core skills sans AI)

Do:

• Copyright considerations
• Privacy considerations
• Misinformation
• AI reliance - test yourself!

Reflection Questions:

1. What assumptions am I making when using this tool?
2. Does this tool present information from the actors I'm interested in learning from?
3. Does this tool allow me to outsource processes that help me learn?
4. Is information being presented in a way that might be susceptible to bias or unfairness?
5. How do the people behind this tool make money? What does this mean for users?

Thank You!

Come say hi!

HIL 116 or catherine.gracey@unb.ca