On May 6th and 7th, I attended the Manitoba Libraries Conference hosted by the Manitoba Library Association (MLA). As an MLA member, I deeply appreciate this gathering. It's an invaluable opportunity to reconnect with colleagues, discover innovative practices, and reflect on my own professional growth. This year was especially exciting as I co-presented a session titled "Demystifying ChatGPT: AI Innovations for Libraries & Digital Repositories" alongside Mike Ellis.

​The conference began with a powerful keynote by Niigaan Sinclair. Niigaan, an Anishinaabe professor from Peguis First Nation, immediately captured my attention with his compelling storytelling and incisive commentary. He contextualized Manitoba’s past and present, thoughtfully reflecting on the Legislative building and the statues toppled in recent years. His point about the absence of Indigenous representation being akin to starting a story at chapter two was particularly impactful. Niigan’s discussion on generational change, highlighted by Manitoba electing Canada’s first Indigenous premier, Wag Kinew, provided insight and perspective. His masterful balance between serious topics, such as residential schools and red dress day, and his use of humour underscored the value of open and straightforward conversations. 

​The first session I attended, “Not Just for Kids: Engaging Adults and Building Community Using Storytime and Music Programs”, led by Austin Matheson and Brittany Lagasse from Winnipeg Public Library, was delightful. It expanded my perspective on adult programming and reminded me of the potential for community-building through creative initiatives like ukulele jams. Given that my previous assistant branch head occasionally serenaded us with her ukulele, this session triggered some memories. 

​After preparing the laptop for my presentation, I quickly assembled a delicious lunch plate, though I had to temporarily stash it behind the projector screen. Despite starting slightly late due to the lunchtime rush, Mike and I had an impressive turnout, with attendees overflowing onto the floor! Mike’s engaging case study on PastFORWARD, Winnipeg Public Library’s digital repository, showcased an innovative AI application in archiving and elicited both laughter and lively participation from the audience. Although time for questions was limited, attendees raised insightful queries about generative AI trained on creative commons materials and the environmental implications of AI. Post-session, I enjoyed meaningful one-on-one discussions about generative AI and potential applications, including possibilities for interlibrary loan systems. 

Afterward, I enjoyed my lunch in the main hall and had an engaging conversation with Trevor, a new connection who shared interests in generative AI, libraries, astronomy, camping, and world travel.

​The afternoon continued with enlightening lightning talks on diverse library initiatives, from updating furniture (“Hold on to Your Seat - Or Don't!”) to enhancing bilingual collections and supporting male caregivers in early literacy programs. These brief yet impactful presentations sparked numerous programming ideas for my own library.

​The day concluded wonderfully with finger foods and mingling, leaving me eager for day two.

​Day two started with the MLA Annual General Meeting, providing a relaxed and productive beginning to the day. It was wonderful connecting with colleagues over coffee, meeting new faces, and exchanging insights.

​The first session of the day, “Nature Programs in a Rural Public Library: Hatching Chicks and Growing Vegetables”, inspired fresh ideas for nature-focused programming. Learning about initiatives like donating produce grown in library gardens reinforced the innovative ways libraries serve their communities.

​“Staff Picks: A Fun, Online Readers’ Advisory Program Model for Your Library” provided practical inspiration for an upcoming autumn ‘Staff Picks’ display. A valuable takeaway from this session was the reminder that “tech should be a tool that supports what you do, not dictate it.”

​The session “In Search of the Lost Library”, presented by librarians from the University of Winnipeg, demonstrated creative solutions for addressing discrepancies in catalogue entries. While their final solution didn’t utilize generative AI, their recognition of it as a potential solution brought a smile to my face, aligning with my interest in integrating AI into library workflows.

​Lunch and the awards ceremony, featuring speaker Chimwemwe Undi, were enjoyable and celebratory. Congratulations to all award winners!

​In the afternoon, Sam Popowich’s session, “The Cultural Politics of Libraries”, was particularly thought-provoking. Sam compellingly argued for recognizing libraries as politically active institutions, examining the 'enlightenment' versus 'social control' perspectives on library history. After the session, Sam generously gifted me his book, "Solving Names: Worldliness and Metaphysics in Librarianship", a thoughtful gesture and a read I’ve already begun to enjoy. 

​The final session I attended, “The Burnt-Out Librarian: Moving on From Vocational Awe”, tackled an important yet often overlooked issue. Carolyn and Monique shared personal experiences and offered practical strategies to address burnout, reinforcing the importance of maintaining healthy engagement with our profession.

​If you’re interested in exploring the content from my session, I’ve included two versions of the presentation slides in PDF format: a short presentation version (as delivered at the conference) and a more detailed version for deeper context and explanation. I hope these resources offer insight into our session and inspire new ways to explore the role of generative AI in libraries. 

​Reflecting on these two enriching days, I felt a great sense of community and connection. The Manitoba Libraries Conference reaffirmed my passion for librarianship, highlighted extraordinary work happening throughout Manitoba, and reinforced my belief that librarians and library workers truly do rule.

The digital world is brimming with information—but not all of it is accurate. With AI-generated content flooding our feeds and misinformation becoming more sophisticated, verifying facts has never been more crucial. Whether you're researching for work, keeping up with the news, or simply scrolling through social media, sharpening your fact-checking skills can help you separate truth from deception. Here’s how to sharpen your perception, enhance your awareness, and become a more informed consumer of information. 

1. Research the Author or Organization
A source’s credibility matters. Before trusting information, investigate who is behind it:

• Is the author a subject matter expert?
• Is the organization funded by groups or individuals who could bias their reporting?
• Does the author or organization have a history of accuracy?
• Are they affiliated with a reputable institution?
• Do they have an agenda?

Looking beyond an entity’s website—by checking independent reviews and other sources—can provide a clearer picture by revealing underlying motives and biases.

2. Use Smart Search Techniques
Finding reliable sources quickly depends on how you search. Here are a few techniques to refine your results:

• Use quotation marks (e.g. "digital literacy initiatives") to search for exact phrases.
• Use site:[URL] (e.g. site:ala.org digital literacy) to search within a specific website.
• Use the minus sign (e.g. -fake) to exclude unwanted terms.
• Conduct reverse image searches to verify photos.

3. Verify the Original Source
Many articles cite secondhand sources—but are they trustworthy?

• Track down the original source or publication.
• Check if the claim is accurately represented and contextually sound.
• Check if multiple reputable sources report the same facts.

4. Consult Fact-Checking Websites
Independent fact-checkers help cut through the noise. Some recommended resources include:

• General: Snopes, PolitiFact, FactCheck.org
• Scientific: SciCheck, Science Feedback
• Canadian-Specific: CBC News, MediaSmarts

​5. Pause and Reflect
If a claim sparks an emotional reaction, that’s a red flag. Misinformation thrives on outrage and urgency. Before sharing or believing a story, take a step back and ask:

• Does this seem too good (or too bad) to be true?
• Who benefits if this claim spreads?
• Does this seem intentionally provocative or overly simplistic?
• Have multiple reliable sources verified this?

The SIFT Method: A Fast, Effective Approach
Mike Caulfield’s SIFT method offers a quick way to assess information:

• Stop – Before engaging, consider the source’s credibility.
• Investigate – Look into the author, website, and their reputation.
• Find Better Coverage – Seek out high-quality sources for confirmation.
• Trace the Claim – Follow information back to its original context.

The P.R.O.V.E.N. Method: A Deeper Dive
For more thorough evaluation, use the P.R.O.V.E.N. method:

• Purpose: Understand why the information was created.
• Relevance: Check if the information is applicable. 
• Objectivity: Assess if multiple viewpoints are included. 
• Verifiability: Confirm the information through reliable independent sources.
• Expertise: Confirm the author’s qualifications. 
• Newness: Determine if the information is current. 

Lateral Reading: Thinking Like a Fact-Checker
Instead of staying on one page, open new tabs and check:

• Wikipedia for background on an organization or expert.
• News archives for related reporting.
• Reviews and expert opinions from multiple sources.

​AI-generated content isn’t going anywhere, and misinformation continues to evolve. Developing strong fact-checking habits keeps you informed and protects you from misleading claims.

​Next time you come across a viral story, a surprising statistic, or a claim that seems off, take a moment to verify before you share. The more we question, the better we can navigate today’s information landscape.

​Have a favourite strategy or a trusted source you rely on? Let’s discuss in the comments!

I've always been curious about how well ChatGPT can generate functional code. To test its capabilities, I decided to start with something relatively simple but still interactive: coding a basic Space Invaders game. I wanted to see how well ChatGPT could generate a working program, how adaptable it would be to my requests, and whether I could refine and improve the code through iterative prompts. This experience turned into an engaging coding experiment, showing me just how powerful AI-assisted development can be.

​Space Invaders is a classic arcade game where the player controls a spaceship that moves left and right, shooting enemies descending from the top of the screen. The game involves essential programming concepts like:

• Rendering graphics: Drawing the spaceship, enemies, and bullets on the screen.
• User input: Allowing the player to move and fire using keyboard controls.
• Game logic: Detecting collisions between bullets and enemies, tracking scores, and resetting the game when necessary.
• Dynamic difficulty: Adjusting enemy speed to increase or decrease the challenge.

To keep things lightweight and browser-based, I opted to use CodeSkulptor, a Python environment designed for educational programming that runs entirely in the browser and uses the simplegui library for handling graphics and game loops.

I began by asking ChatGPT to generate a basic Space Invaders game in CodeSkulptor. The initial version included:

• A rectangular player-controlled spaceship.
• Simple movement using the left and right arrow keys.
• Enemies that moved downward at a fixed speed.
• A basic shooting mechanism where bullets fired straight up and disappeared when hitting enemies or leaving the screen.

However, I quickly realized some limitations and requested modifications:

1. Increased Playability: The original screen size was too small, and the player’s speed felt sluggish. I asked ChatGPT to triple the spaceship’s speed and expand the game window to provide more space for movement.
2. Balanced Enemy Movement: The enemies were initially too fast, making the game too difficult. I asked for their speed to be reduced to one-third of the original rate.
3. Dynamic Difficulty: I requested that enemy speed increase by 5% when pressing the up arrow key and decrease by 5% when pressing the down arrow key. This allowed for on-the-fly difficulty adjustments.
4. Visual Improvement: I wanted the spaceship to be a triangle rather than a rectangle to better represent a traditional spacecraft.
5. Game Reset Feature: Finally, I asked for a way to restart the game with a key press. ChatGPT implemented a reset function triggered by pressing the 'N' key, which cleared the screen, reset the score, and respawned enemies at random positions.

​By the end of this experiment, I had a functional and customizable Space Invaders game running in CodeSkulptor. The iterative process demonstrated how well ChatGPT can understand and implement coding requests, allowing for quick modifications and enhancements.

For those interested, I highly recommend trying out the code in CodeSkulptor and playing around with your own modifications. The experience is a great way to learn how game logic works while also exploring AI-assisted development.

This experience left me excited to continue exploring coding with ChatGPT. Now that o3-mini-high is available—a model that is supposed to be even more proficient at coding—I’m even more eager to see how it improves code generation, debugging, and refactoring. With each iteration, AI models are becoming more adept at understanding context, implementing changes effectively, and even suggesting improvements I might not have thought of myself.

​If this simple Space Invaders game was just the beginning, I can only imagine how far AI-powered coding assistance can take us. Imagine being able to code entirely in natural language! Whether you're a beginner learning the basics or an experienced developer looking to prototype ideas quickly, ChatGPT is proving to be an invaluable tool in the coding process.

I've been playing around with ChatGPT for a while now, experimenting with its ability to generate and refine stories, especially those rooted in science. One of my latest projects was crafting a science fiction short story that balances scientific accuracy with a sense of curiosity and wonder—something in the flavour of Carl Sagan.

The process? A mix of notes, structured planning, AI-assisted brainstorming, research, and a lot of tweaking:

• Story Concept & Playtime: I started with a basic idea—a Type Ia supernova unfolding across different cosmic distances—and played around with ChatGPT, bouncing ideas back and forth.
• Academic Meets Creative: I had ChatGPT analyze and draw from one of my old university astronomy papers to reinforce the scientific foundation of the story.
• Outlining the Narrative: I structured the story into five chapters, each set at a different vantage point, from the doomed white dwarf, to an autonomous research probe, to a crewed space ship, to the Earth, to the Andromeda Galaxy. One focus was scientific knowledge, but another was evoking that deep sense of cosmic wonder.
• Experimenting with ChatGPT Models: I tested different inputs, played around with phrasing, and fine-tuned the best output before mixing in my own edits.

To enhance the experience, I experimented with Sora to create short videos for each chapter, the title screen, and ending. I compiled these into videos for the story, trying two different approaches:

1. Text-Based Generation: I copy-pasted the chapter text directly into Sora and went with whatever it generated.
2. Prompt-Based Generation: I asked ChatGPT o3-mini to come up with suitable Sora prompts—one for each chapter, the title screen, and ending. I then put those prompts into Sora with minimal editing. This worked somewhat better but Sora still had issues generating text correctly.

Both videos turned out alright, but still left much to be desired. Sora has potential, but there’s plenty of room for improvement—at least for us humble ChatGPT Plus users.

On top of that, I wanted an atmospheric soundtrack, so I used ChatGPT to craft a dungeon synth instrumental prompt for Suno. I’ve been really into dungeon synth lately, and this story felt like the perfect inspiration for something melancholic, immersive, and cosmic. 

Suno generated two versions:

• Luminous Shadows - Version 1

​After all that experimenting, refining, and assembling, here’s the final outcome: ‘Echoes of a Dying Star’—a story that explores the cosmic scale of a supernova through the perspectives of a doomed autonomous research probe, a distant spaceship, Earth-based observers, and even the Andromeda Galaxy.

Check out the story, and watch the accompanying videos below!

First, I experimented with ChatGPT 4, using the following prompt:

     Write a fairytale in the style of the Brothers Grimm to communicate the         meaning of the following quote: “insert quote here.”

​I then tested ChatGPT o1 with the same prompt. 

With this straightforward instruction, I watched as ChatGPT spun imaginative fairy-tale narratives, each crafted to illuminate the core message of the quote in question. Despite the simplicity of the prompt, the stories that emerged were rich in imagery and moral depth, perfectly capturing that timeless Brothers Grimm flair. ChatGPT 4 tended to produce more intricate plots with rich symbolism and nuanced character development, while ChatGPT o1 tended to generate simpler, more straightforward tales that still conveyed the meaning of quotes effectively. Both versions demonstrated a strong ability to distill meaning into compelling folklore-style narratives.

One of the things that impressed me most was how easy it would be to adapt these short stories for story time programming at public libraries. Librarians can quickly generate age-appropriate tales with built-in morals, lessons, or pop culture phrases, then share them with young audiences in a fun, engaging way. And if you add an image generation tool—like DALL·E or another AI art generator—you can produce simple illustrations to accompany each story, making the entire experience even more magical and visually appealing.

​There’s real value in these stories, not just for children but for adults as well. Narratives are a powerful way to communicate values, spark imagination, and provide insight. They allow us to see ourselves in characters and situations, giving life to abstract ideas in a relatable form. Whether you’re reinforcing a personal mantra, teaching a life lesson, or simply entertaining an audience, stories transcend age and background. They capture the heart and the mind all at once.

​For those looking to explore creative writing or enhance their library programs, I highly recommend experimenting with ChatGPT. It’s an enjoyable and effective way to bring fresh, imaginative content to life—whether you’re spinning your own fairy-tale world or passing on time-honoured wisdom in new and exciting forms.

​In today's fast-paced digital world, leveraging artificial intelligence (AI) can significantly enhance productivity and efficiency. ChatGPT, developed by OpenAI, is a powerful language model that can assist with a wide range of tasks, from drafting emails to brainstorming ideas. This guide explores the philosophy behind integrating ChatGPT into your workflow and provides practical steps on how to incorporate it effectively.

​Enhancing Productivity and Efficiency
At its core, ChatGPT is designed to augment human capabilities. By handling routine or time-consuming tasks, it allows you to focus on more complex and creative aspects of your work.

​Complementing Human Skills
ChatGPT excels at processing and generating text based on patterns in data it was trained on. While it doesn't possess consciousness or personal experience, it can provide insights and suggestions that complement human intuition and expertise.

​Continuous Learning and Adaptation
Incorporating ChatGPT into your workflow is not a one-time setup but an ongoing process. As you interact with the model, you learn how to craft better prompts, and the AI can provide more tailored responses over time.

1. Identify Suitable Tasks
Determine which aspects of your workflow can benefit from AI assistance. Common areas include:

• Content Creation: Drafting articles, emails, social media posts, and image generation.
• Brainstorming: Generating ideas or exploring different angles on a topic.
• Research: Summarizing information or explaining complex concepts.

2. Set Up Access
Access ChatGPT through platforms that offer the service, such as:

• OpenAI's ChatGPT Website: Interact directly via the web interface.
• OpenAI's ChatGPT App: Use the official ChatGPT mobile app for iOS and Android, or the desktop application for convenient access and functionality.
• Integrations: Use plugins or extensions that incorporate ChatGPT into applications like Chrome or Microsoft Teams.
• API Access: For developers, integrate ChatGPT into custom applications using the OpenAI API.

3. Craft Effective Prompts
The quality of the AI's output depends largely on the input it receives.

• Be Specific: Clearly state what you need. Instead of "Tell me about climate change." try "Summarize the impact of climate change on coastal cities."
• Provide Context: If the task is complex, give background information to guide the AI.
• Set the Tone and Style: Indicate if you prefer a formal report, casual explanation, or bullet-point list.

Planning Healthy Meals
Use ChatGPT to create meal plans or recipes based on ingredients you have or dietary preferences.

• Example Prompt: "Help me create a healthy dinner using salmon, asparagus, and brown rice. I prefer low-carb options."

Developing Meal Plans
Ask ChatGPT to generate a week's worth of healthy meal ideas.

• Example Prompt: "Plan a 5-day vegetarian meal plan for me, including breakfast, lunch, and dinner. I need quick recipes that take under 30 minutes to make."

Planning a Trip
Utilize ChatGPT to help organize your travel itinerary, find attractions, or suggest packing lists.

• Example Prompt: "I'm planning a 7-day trip to New Zealand. Suggest an itinerary that includes hiking, local culture, and must-see landmarks."

Minimizing and Organizing
Get advice on decluttering and minimizing possessions.

• Example Prompt: "Provide a step-by-step plan to declutter my closet and minimize my wardrobe to essentials."

Time Management
Improve your productivity by optimizing your schedule.

• Example Prompt: "Help me create a daily routine that balances work, exercise, family time, and personal hobbies like reading."

Assisting with Reference Questions
Library staff can use ChatGPT to help answer patron inquiries more efficiently, including reader's advisory and identifying books based on limited information.

• Example Prompts:
  • "Provide a simple explanation of the expansion of the universe suitable for a middle school student."
  • "A patron remembers a mystery novel about a detective who solves crimes in his dreams, and the book cover features a clock melting. Help identify the book title."

​Generating Image Ideas and Descriptions
Use ChatGPT to help conceptualize images or create descriptions for image generation tools.

• Example Prompt: "Describe an image of a public library hosting a community cooking class, with people of all ages preparing meals together."

Creating Prompts for AI Image Generators
Craft detailed prompts to be used with AI image generation models like DALL·E.

• Example Prompt: "Help me write a prompt for an AI image generator to create an illustration of a minimalist workspace with a laptop, notebook, and a cup of green tea by a window overlooking a forest."

Generating Event Ideas
Brainstorm new programs to engage the community.

• Example Prompt: "Suggest five interactive workshop ideas for adults at a public library focusing on digital literacy."

Understanding Complex Topics
Leverage ChatGPT to break down complicated subjects and aid in your personal learning journey.

• Example Prompts:
  • "Explain the process of star formation and the lifecycle of stars in simple terms."
  • "What is the significance of exoplanet discovery in our understanding of the universe?"

Be Clear and Specific
The more precise your prompt, the more relevant the response.

• Ineffective Prompt: "I need meal ideas."
• Effective Prompt: "Provide three healthy dinner recipes using chicken, spinach, and sweet potatoes."
• Ineffective Prompt: "Help me plan a trip."
• Effective Prompt: "Create a 4-day itinerary for a budget-friendly trip to Japan, focusing on onsen, nature, culture, and local cuisine."
• Ineffective Prompt (Library Use): "Help with programs."
• Effective Prompt (Library Use): "Outline a plan for a children's reading hour themed around learning the night sky."

Use Iterative Refinement
If the first response isn't quite right, refine your prompt by providing additional details or asking follow-up questions.

• Meal Planning Follow-Up: "These recipes are great, but I have a nut allergy. Can you adjust them accordingly?"
• Trip Planning Follow-Up: "Can you include some lesser-known attractions and recommend affordable accommodations?"
• Minimizing Stuff Follow-Up: "I have a lot of books. How can I decide which ones to keep and which to donate?"
• Library Programs Follow-Up: "Great ideas! Can you include suggestions on how to promote these workshops to attract more participants?"

Leverage for Learning
Use interactions as a learning opportunity to deepen your understanding.

• Example Prompts:
  • "What are the basic steps to start learning a new language effectively?"
  • "List some beginner-friendly astronomy projects I can do at home to learn about and memorize constellations."

​Avoiding Over-Reliance

• Supplement, Don't Replace: Use ChatGPT to augment your work, not replace critical thinking or professional expertise.
• Continuous Learning: Stay informed and keep developing your skills beyond AI assistance.

​Integrating ChatGPT into your workflow can be a game-changer, enhancing efficiency and sparking creativity. By understanding its capabilities and limitations, and applying best practices, you can harness this powerful tool to complement your skills and advance your projects. Embrace the collaborative potential of AI, and let it be a catalyst for innovation in your personal and professional journey.

​Remember: The key to effectively using ChatGPT is experimentation and adaptation. As you become more familiar with how the AI responds, you'll discover increasingly valuable ways to incorporate it into your daily tasks.

Happy New Year, everyone! I hope you’re all kicking off 2025 feeling refreshed and ready to discover new stories. Looking back on 2024, I can’t help but smile at the incredible books I got through this past year. It’s always tough to narrow things down, but here are a few standouts I couldn’t put down.

‘Machine Vendetta’ by Alastair Reynolds
This third installment in the Prefect Dreyfus Emergency series hit all the right notes. Once again, Reynolds brilliantly combines intricate mystery with a universe so vivid you can practically feel the vacuum of space. From advanced lemur-like beings to AIs vying for control, the story is both a thought-provoking investigation and a stunning exploration of what humanity might become.

‘Berserk’ Series (Deluxe Edition) by Kentaro Miura, Translated by Duane Johnson
While I read most of the series in previous years, I finished the series this year. The final Deluxe Edition volume was bittersweet. The abrupt ending, a result of Miura’s untimely passing, made the included guidebook feel extra special. It’s a fitting tribute to Miura’s extraordinary world—an epic adventure both beautiful and brutal, packed with unforgettable characters and jaw-dropping art.

‘Architects of Intelligence: The truth about AI from the people building it’ by Martin Ford
This collection of interviews with AI pioneers is a must-read if you’re curious about where technology is headed. Ford steers insightful conversations with leading minds, making dense concepts accessible. You’ll find yourself pondering everything from ethics to existential risks, all while gaining a richer understanding of the true potential (and pitfalls) of AI.

‘Schismatrix Plus’ by Bruce Sterling
Sterling’s vision of a posthuman future brims with creativity. The clash between the biologically enhanced Shapers and the cybernetically augmented Mechanists sets the stage for complex social dynamics and wild philosophical questions. Featuring “Lobsters” that push the boundaries of what it means to be human, this book still has me thinking about the essence of identity and evolution.

These were just the highlights. Other favourites that made 2024 unforgettable include Ray Kurzweil’s ‘The Singularity Is Nearer: When We Merge with AI’, Rob Guillory’s ‘Farmhand’ series, Neil Gaiman’s ‘Norse Mythology’, and the four-volume ‘Aâma’ series by Frederik Peeters. Each one left its mark on my reading year.

If you’d like to connect, feel free to send me a friend request on Goodreads. Let’s make 2025 a year of great reads and recommendations. Here’s to a fantastic year ahead!

​In this final post of our series, I’m excited to take you on a journey beyond the standard use of AI and into the realm of creativity and specialization. AI isn’t just about answering questions—it’s a tool for sparking creativity, simulating professional consultations, and generating unique content. Let’s explore how creative and specialized prompting techniques can expand your horizons and help you get the most from your AI interactions.

1. Scenario-Based Prompting

• Definition: Asking the AI to create scenarios to visualize outcomes or explore hypothetical situations.
• Example: "Imagine a library of the future where all services are delivered through virtual reality. Describe a day in the life of a library user in this setting."
• Best For: Storytelling, creative writing, speculative fiction, or exploring future possibilities.

This technique is ideal for creative writers or futurists who want to explore "what if" scenarios. It allows for imaginative and innovative responses that help visualize potential futures or creative storylines.

2. Comparative Prompting

• Definition: Requesting the AI to compare and contrast different concepts, ideas, or entities.
• Example: "Compare the advantages and disadvantages of offering eBooks vs. physical books in a public library."
• Best For: Highlighting distinctions, analyzing choices, or presenting balanced viewpoints.

Comparative prompting is particularly useful for analysis and decision-making processes. By comparing different options, you can generate balanced and well-rounded discussions.

3. Image Generation Prompting

• Definition: Using vivid, descriptive language to guide AI in generating images.
• Examples: 
  • "Create an image of a bustling market in a futuristic city, with diverse alien species trading goods under glowing neon lights."
  • "Create an image of a vibrant children's storytime session in a multicultural library, with diverse children, colourful decorations, and engaging storytelling props."
• Best For: Visual content creation, generating stock images, conceptual design, or artistic inspiration.
• Tips:
  • Use Specific Descriptors: Adjectives like "serene," "vibrant," "foggy," or "chaotic" help in crafting precise visuals.
  • Include Elements and Themes: Mention key components (e.g. "a spaceship landing", "ancient ruins in a jungle", or “a rustic barn on the prairie”).

For artists and designers, image generation prompting provides a powerful tool to visualize concepts. Using detailed language ensures that the AI captures the intended mood and details.

4. Hybrid Prompting

• Definition: Combining different prompting methods, like few-shot and chain-of-thought, for multi-layered responses.
• Examples: 
  • "Provide three examples of how libraries support lifelong learning. Then, outline a step-by-step approach for developing a new adult education program focused on career skills."
  • "Provide three examples of how AI is transforming customer service. Then, outline a step-by-step approach for implementing AI-driven customer service in a retail company."
• Best For: Tackling multifaceted topics, generating comprehensive guides, or combining diverse information sources.

Hybrid prompting offers the best of both worlds by blending different techniques to achieve complex and nuanced outputs. It's ideal for comprehensive content that requires multiple layers of information.

5. Personalized Prompting

• Definition: Tailoring prompts to user preferences or needs to create more customized content.
• Examples: 
  • “I am interested in hard science fiction, space operas, transhumanism, and intricate worldbuilding. Please recommend ten novels or series that strongly incorporate these themes.”
  • "Based on interests in historical fiction, mystery novels, and outdoor adventure, recommend a weekend itinerary that includes related activities."
• Best For: Providing highly relevant suggestions, recommendations, or plans based on individual preferences.

Personalized prompting makes AI interactions more relevant and engaging. By focusing on user-specific needs, this technique ensures the content is highly tailored and practical.

​And that concludes our exploration of AI prompting techniques! From building a solid foundation to unlocking deeper insights and expanding your creative potential, you now have a robust toolkit for engaging with AI. Whether you're using these techniques in a library setting, for research, or to fuel creative projects, remember that the key to mastering AI lies in thoughtful, intentional prompting. Thank you for joining me on this journey—let's continue exploring the power of technology and the limitless potential it offers.

​Welcome back! As we continue this series on structuring AI prompts, it’s time to build upon the foundational techniques we’ve already explored. In this post, we’ll learn advanced prompting techniques that can help you draw out deeper insights and more sophisticated responses from AI. Whether you’re working in a research-heavy field or diving into a creative project, these advanced techniques will help you go further.

1. Chain-of-Thought Prompting

• Definition: Guiding the AI to articulate a logical, step-by-step breakdown of a complex idea or process.
• Example: "Outline the steps to develop a successful early literacy program at a local library."
• Best For: Explaining multi-step processes, enhancing logical reasoning, or providing educational content.

Chain-of-thought prompting is particularly useful for educators or anyone looking to break down complex ideas into manageable steps. It enables the AI to provide a more structured and logical explanation, ideal for tutorials and how-to guides.

Framework Tip: Use the C-A-R-E framework:

• Context: Provide the background for the library.
• Action: Describe the steps involved.
• Results: Clarify what the outcome should look like.
• Examples: Give examples for each step.

This structured approach helps AI break down processes systematically, making it ideal for educators and professionals who need detailed explanations.

​2. Role-Specific Prompting

• Definition: Instructing the AI to respond from the perspective of a specific profession or expertise.
• Example: "You are a library technology specialist. Provide a strategy for implementing a digital literacy workshop series for seniors."
• Best For: Generating expert-like advice, simulating consultations, or producing specialized content.

By assigning a role to the AI, you can tailor responses to match specific professional contexts, making this technique valuable for content that demands a high degree of expertise or specialization.

Framework Tip: Utilize the R-T-F framework:

• Role: Act as a [Role].
• Task: Create a [Task].
• Format: Show as [Format].

This technique allows users to simulate consultations or expert advice, making it highly relevant for creating specialized and professional content.

3. Prompt Chaining

• Definition: Dividing complex topics into smaller, connected prompts that build upon each other for comprehensive exploration.
• Examples: 
  • "List the key components of a library's collection development policy."
  • "Describe how user feedback can influence collection development decisions."
  • "Provide examples of successful collection development strategies from public libraries."
• Best For: Exploring topics in detail, creating structured and in-depth discussions, or educational tutorials.

Prompt chaining is great for diving deep into a subject by breaking it down into interconnected questions. This method ensures a thorough exploration and helps build a complete picture of a topic.

Framework Tip: Apply B-A-B:

• Before: Explain the problem or situation.
• Action: State the goal or desired action.
• Output: Ask for a specific format.

Framework Tip: Use the A-P-E framework to control the output:

• Action: State the task.
• Purpose: Create a purpose for the response.
• Expectation: Describe what should be avoided.

Negative prompting helps AI think outside the box and avoid common responses, ideal for generating unique and thought-provoking content.

​Advanced prompting techniques open doors to deeper, more insightful AI interactions. By applying these strategies, you can get more nuanced and detailed responses that meet your professional or creative needs. In the final post of this series, we’ll explore creative and specialized prompting techniques that push the boundaries of AI’s capabilities, allowing for more imaginative and customized interactions. Stay curious!

Welcome to the start of an exciting journey into the world of AI interactions. As someone who has spent years merging the worlds of libraries, philosophy, and technology, I’ve seen how AI tools like ChatGPT can transform the way we gather and use information. This three-part series will guide you through core, advanced, and creative techniques for structuring AI prompts to maximize the usefulness of AI in your professional or creative pursuits. Whether you're a librarian looking to enhance reference services or someone eager to experiment with AI's creative potential, these techniques will help you get the most out of your interactions.

​To interact effectively with AI, it’s crucial to start with a strong foundation. In this first post, we’ll explore core prompting techniques. Core prompting techniques help establish the basics of AI communication, ensuring clarity and precision in responses. Let’s dive in.

1. Zero-Shot Prompting

• Definition: Directly asking the AI to perform a task without providing any prior examples.
• Example: "Explain the concept of digital literacy."
• Best For: Simple definitions, explanations, or introducing new topics.

Zero-shot prompting is like starting with a blank slate. It's perfect for straightforward queries where you're looking for concise answers or explanations. When you're unsure of the AI's capacity or want a broad overview of a subject, zero-shot prompts are the go-to technique.

Tip: Specify the tone (e.g. formal, informative) and format (e.g. bullet points) to refine the response. For example, "Explain the key concepts of digital literacy in a concise, bullet-point format." This makes the response more tailored and specific to your needs.

2. One-Shot Prompting

• Definition: Providing one example to guide the AI in crafting a response.
• Example: "Describe the concept of 'resilience' in leadership, giving one example from a well-known leader."
• Best For: Clarifying specific concepts or terms, offering precise guidance through a single example.

One-shot prompting helps narrow down the focus. By providing a single example, you guide the AI toward the desired direction, ensuring more accurate and context-aware responses. It's particularly useful for slightly more nuanced topics where you need the AI to consider a specific approach or angle.

Tip: Define the audience (e.g. leadership trainees) and use analogies to make the explanation more relatable. For instance, "Describe 'resilience' in leadership for new managers using an example of a famous leader." This helps in aligning the response with the audience's understanding.

3. Few-Shot Prompting

• Definition: Offering 2-5 examples to help the AI identify patterns and provide richer responses.
• Example: "Give three different scenarios where augmented reality can enhance education."
• Best For: Achieving a more nuanced understanding and exploring diverse applications of a topic or concept.

When you need depth and variety, few-shot prompting is the answer. Providing multiple examples allows the AI to learn patterns and apply a broader range of knowledge, resulting in more comprehensive responses.

Tip: Include context or terminology relevant to the field (e.g. "augmented reality in education") to narrow down the scope and focus the AI's response on relevant details.

4. Iterative Prompting

• Definition: Asking follow-up questions or refining initial prompts based on previous AI responses.
• Example: "What are the benefits of offering maker spaces in libraries? Now, focus specifically on how they support STEM education."
• Best For: Diving deeper into specific aspects of a topic, refining the focus to enhance the depth of responses.

Iterative prompting allows for dynamic and evolving conversations with the AI. By building on previous responses, you can explore a topic in detail, uncovering various layers and perspectives.

Tip: Use questions to drive deeper exploration and ask the AI to consider multiple perspectives or counterarguments for balanced insights. This technique is ideal for complex topics that require thorough exploration from various angles.

​By mastering these core prompting techniques, you’ve taken your first step toward building more meaningful and accurate interactions with AI. Whether you’re crafting simple inquiries or laying the groundwork for deeper exploration, these basics will serve as your compass. In the next post, we’ll dive deeper into advanced prompting strategies that unlock the true potential of AI in professional contexts. Stay tuned for more!