Academics

This month I’ve learned a lot about the Odyssey Program and developed a lot of plans. The first thing we worked on were our elevator pitches, which were descriptions of our projects with some personal details. We worked on them extensively and practiced them many times. Eventually, we presented them to a gathering of mostly family and friends. It was helpful to reacquaint myself with public speaking and I believe it went well.

After we had completed our elevator pitches and presentations we moved on to planning and research. I developed a detailed plan for this year and what I’d like to accomplish each month. We went in-depth on types of sources and some good places to use for research like JSTOR, a digital database of articles and studies. I acquired four sources relating to my project and drafted an email to Tom Scott about a possible interview. Overall I am happy with the progress I’ve made so far and am looking forward to learning more about my field.

After all the commencement activities I was able to really start on my project and research relevant topics. As of writing, I have read and annotated three major papers. One was a research report on the product landscape for AI in higher education, a quote of particular interest being “If current trends continue, in the near future GAI use will be ubiquitous, fully integrated into the core mission of colleges and universities”. Another was study about who is most exposed to AI, which while helpful, was more of a jumping off point for other more targeted sources. Specifically, the report put me onto a Harvard study researching how AI might be best implemented into everyday life. It concluded that it will be most important to identify which niches AI best fits into and build systems around them. I was reassured to learn that answering course specific questions and processing documents was among the AI’s capabilities.

The study concluded that it will be most important to identify which niches AI best fits into and build systems around them. I was reassured to learn that answering course-specific questions and processing documents were among AI’s capabilities. Overall, I’m better researched and more confident than last month, and I am looking forward to working with the ChatGPT API.

I have done extensive research on more sources related to my project and found many that will be helpful in my eventual implementation. However, it has been difficult to find new sources that will actually be valuable. My field is a new and emerging one which means most research is still ongoing and lots of information is still unavailable to the public. I am hopeful that once I have completed my research requirements, I will be able to launch into working with the ChatGPT API and gain some insight into how to work with a more raw version of AI.

On the business side, I've done a lot of work, including our mission statement which I feel is relevant enough to share here “Our mission is to integrate artificial intelligence into industries to empower organizations, enable their workforce, and foster efficiency. We aim to create AI solutions that benefit industries while making work easier and more impactful for employees.”

Throughout this semester I have made great progress in learning about AI. I began with the elevator pitch and research question, both of which helped to clarify what my project was about. Through it, I realized that I would develop an AI teacher’s assistant that could create and grade assignments, gather feedback, answer student questions, and more, all tailored to the teacher, who provides the AI with information, allowing it to learn and adapt to their unique style. The Aim, of course, would be to allow teachers to spend more time with students, not to replace them. It would also be a huge help to students, allowing them to see more up-to-date information about their grades, and ask questions specific to the classes they’re in twenty-four-seven. Through developing this AI I’d learn about machine learning, coding, and a plethora of other things I don’t even know about yet starting with Google’s AI Essentials course.

Google’s course ended up being my community observation. I completed the first three modules for Google’s “Generative AI for Developers Learning Path,” and I hope to complete one each semester, ending the final semester with its completion. I found the concepts extremely helpful, especially the third module. The first module talked in broad terms about how an AI model interprets images, generates them, and the training process that improves its capabilities. Next, the second module talked about how attention mechanisms work, and how sometimes it’s necessary to focus on some words more than others. Finally, the third module focused on encoder-decoder processes and included code that revealed what the models are doing on a base level. I learned a lot of useful information about how AI works in a tangible sense and how I can train my own models. I also learned about useful resources for creating AI models which I anticipate using frequently. Despite its helpfulness, I feel that I won’t be able to get everything I can out of it until I am working with the resources they describe, and I think I will have to review the material when I use it. Overall I feel that this course will be a fantastic resource for my project and I’ll use it not only as a technical guide for specific functions but as a more general resource that can apply to many aspects of my project.

The elevator pitch not only clarified things but gave me experience working with the other scholars and speaking in a public-ish setting. From there I began work on my annotated bibliography which proved to be challenging because of its specificity and the work required to make it. I continued working on my annotated bibliography and learned a lot about how AI is used and viewed in the world. I learned about how there are different classifications for AI-human collaborations, the most effective being centaurs. Centaurs refer to humans being supported by AI but still leading. I also learned about the jagged technological capabilities of AI, which refers to the random-seeming abilities of AI, which are concentrated in certain areas and nonexistent in others.

The annotated bibliography culminated in my literature review, in which I concluded that together the sources identify how transformative AI has the potential to be across most sectors, and most importantly for my project, education. Some of the key findings include the mass adoption of AI services in education and its ability to improve both efficiency and quality of outcome. In "Generative AI in Higher Education" and "Artificial Intelligence Applications to Support K–12 Teachers and Teaching," the importance of integrating AI into educational frameworks is a central theme, focusing on tools that assist rather than replace educators. Meanwhile, "Who is Most Exposed to Generative AI?" puts on display how much AI can improve workplace tasks but can cause issues for less skill-intensive jobs. Unique to the previous sources, "Getting AI Right" and "If We Succeed" are ethics-based and call for transparency, diversity in development, and human-AI alignment on values. Originally I didn’t want my project to focus on ethical issues but I now realize that they apply no matter what kind of project you’re working on.

I also conducted an interview with Mr. Aase and gained some insight into how teachers might work with AI. During the interview, we discussed the Applications of AI in education, discussing its strong areas, like math and English for explanation, alternative perspectives, and personalized learning, and its weak areas, like history, where research skills are critical and tasks require personal feedback from teachers. We discussed how AI has been narrow in applications like Turnitin for plagiarism detection, generative AI for project ideas, and early automation tools like Scantron and how it's evolving to be more general and may one day be integrated into general processes. Another topic was how teachers will need to adapt to AI, with the most successful outcomes likely coming from teachers embracing AI while acknowledging boundaries and emphasizing critical thinking with importance placed on teachers developing AI-based skills. During the interview, I gained a lot of insight into how teachers already use AI, and how new tools might best fit into the classroom. Aase described parts of the learning process that should and should not be affected by AI in accordance with both student development and AI capabilities which ties into the jagged technological landscape I’ve been researching.

Broader implications from my findings make clear the advantages and risks associated with AI. I hope to see AI increasing access to education, improving teaching outcomes, and increasing workforce productivity but I am weary of job displacement, and bias. I look forward to applying my knowledge through developing my project.

This month, I’ve begun working with the ChatGPT API, which is how you interact with ChatGPT through code for those who don’t know. I decided that a practice project to learn how the API worked would be a good way to start. I began with the basic ability to generate content from a prompt. The API is surprisingly intuitive, and I learned there are two inputs for the model (technically three, but we’ll get there). There is the system and the user. The system supplies instructions on how to respond, which can include both the AI’s “character” and the literal formatting of the text. For example, the system might instruct the AI to be a pirate and it will respond from that perspective. The other role is user, the normal prompt you enter into the ChatGPT website.

I began by telling the AI that it was a tavern keep, and I entered prompts through the terminal  (a text-based interface that allows users to interact with a program by entering commands while it’s running). From this, I was able to generate prompts in character and was unable to convince the AI to drop its character which was very promising for my project. From there I programmed in local memory to remember the conversation, more characters, a character filing system using JSON, and more. The code looks like this:

I plan to move onto the actual prototype for my final project soon. I also had some extra credits for generating prompts which reset at the beginning of the month, so enjoy some excerpts from my bard character.

Generated Content:

Oh, gather 'round, good folk, come hear this spectral tale,
Of a lonesome, eerie castle where the moonbeams pale.
Upon a hill it stands, cloaked in a misty shroud,
Where ghostly voices whisper, both mournful and loud.

The Castle of Glendenmoor, it's called by the brave,
Yet the tales that are spun could make a knight misbehave.
For within its stony walls, specters roam at will,
Their tales of woe and wonder shall nay be forgotten still.

First, there's Sir Edgar, the knight with a gilded lance,
He lost a tourney fair, and hence haunts in a trance.
Each night he jousts the air, practicing long and hard,
His steed, now but a shadow, rides beside our ghostly bard.

Then Lady Celeste, with her flowing silver hair,
She walks the halls lamenting, with nary a worldly care.
Her heart was once broken by a lord of renown,
Yet she's destined to wander 'til the day he comes 'round.

And let us not forget the cook, old Boris McLean,
Who paces through the kitchen, in a looped refrain.
He seeks the perfect pudding he lost that fateful night,
'Tis said his custard's cursed—so tangy, yet contrite.

The last is the minstrel, a fellow much like me,
He sings a ghostly ballad beneath the willow tree.
His lute is made of echoes and strings of sheer moonlight,
And when his chords do linger, they chase away the fright.

So heed these ghostly tales as the clock ticks midnight near,
Of haunted halls and shadows that float from there to here.
And if you glimpse a glowing knight or hear a spectral tune,
Smile, for they're but stories beneath the silver moon! 🌕

Generated Content:

Oh, gather 'round and hear the tale of a forest steeped in mystic veil,
Where time's own hands choose to dance, in a dreamlike, timeless trance.
The trees are tall with emerald hue, their leaves whisper tales anew,
And in this realm so lush and grand, the moon and sun together stand.

This woodland realm is strangely charmed, where clocks and hours are disarmed,
A minute here feels like a day, and decades whisk in moments' play.
The streams, like threads of silver spun, reflect a sky where two suns run,
While creatures play in vibrant glow, in rhythms only they do know.

The mushrooms giggle as you pass, and squirrels wear a cap of brass,
Fireflies blink in cryptic rhyme, flitting through the folds of time.
A gentle breeze may stop and stare, as if deciding when or where,
To move again and gently sigh, weaving songs that touch the sky.

Should you wander this enchanted wood, where tick and tock are understood,
Be wary of the siren's call, and faeries dancing at nightfall.
For once you leave, though years may seem, mere minutes pass beyond the dream.
Yet tales you'll share of this strange land, where time flows like shifting sand.

Generated Content:

Gather 'round, my friends, for a tale full of dread,  
Of a cursed blade wielded by a lad full of lead.  
It gleamed in the moonlight with a sinister glow,  
A weapon of legend, from long, long ago.

Young Sir Gareth came across it one night,  
In the ruins of a fortress, in dim, gloomy light.  
The hilt was enticing, with gems red as wine,  
But the blade whispered secrets, not quite so benign.

"Draw me forth!" cooed the voice from within,  
"Great power you'll wield, and great battles you'll win."  
With a smirk and a chuckle, Gareth took hold,  
For the tales told by bards had made him bold.

With a swish and a swoosh, he vanquished each foe,
The sword's wicked magic dealt blow after blow.
But soon young Gareth found, to his utter dismay,
The blade craved more than mere men to slay.

Each life it claimed made the shadows grow bright,
Crimson streaks in the sky, a horrible sight.
The sword wanted power, it thirsted for fears,
Drawing life from the wielder through long, weary years.

Gareth felt his spirit begin to unravel,
As the sword led him down a dark, twisty gravel.
To break free from its grip, he sought wisdom and lore,
From witches and wizards and sages of yore.

At last, to a wise druid, he earnestly pled,
"Unshackle my soul from this curse," he said.
With a cauldron a-bubbling and incantations deep,
The druid whispered secrets that made the sword weep.

With a flash and a crash, the cursed bond was no more,
Yet the sword lay lifeless on the cold, stone floor.
Sir Gareth had learned a most valuable thing,
That greed for great power comes with a sting.

So beware, dear brave hearts, of temptations so grand,
For the cost may be more than a mortal can stand.
And remember this tale of a sword and its snare,
When you next hear a whisper, drifting through the air!