What current AI-related issues, developments, or decisions do you find especially relevant to contemporary society? Craft a short post to give your classmates an overview of the issues involved and why it’s so important.

I think one of the most popular issues that are being raised by society is the ethical use of AI in the academic or research field. Nowadays, there are so many AI tools that can be used to boost your productivity, such as ChatGPT, Copilot, and Midjourney. However, the existence of those tools and their use blur the lines around authorship, originality, and intellectual integrity.

In school, with the help of AI, students can now generate a whole essay or a thousand lines of code with just a click. This has raised a serious question about plagiarism, the learning process, or “aigiarism” – the plagiarism using AI. In the workspace, people also raise a debate about whether they should consider AI-generated context as part of the contribution, which could potentially affect the hiring process and performance.

These issues are very important because they create a longstanding assumption about what creativity is. As AI tools are getting easier to access and considering how powerful they are, we urgently need updated policies and education guidelines to ensure that AI can be a righteous tool to enhance human capabilities without replacing human work.

Sources:
https://crossplag.com/what-is-aigiarism/

This week, we encountered a lot of questions about the fairness of AI use in academic fields. As I decided to further my studies in AI, I consider AI to be just a tool for improvement. What I mean by a tool here is that it’s just something similar to Google or Facebook. When Google first came out, the world was blown away: it became a significant moment in the history of the internet. Same as Google, I think Artificial Intelligence is also a revolution in terms of how we access and process information. The key is not whether we should use AI, but how we use it responsibly. To me, AI can enhance learning, boost creativity, and save time when used correctly. It’s not a shortcut to avoid thinking, but a powerful aid to deepen our understanding.

When it comes to determining ” how much is too much” in the use of AI, it is less about the restriction of limits but the intention behind its use. I agree that AI is a great tool in supporting learning, researching, and productivity, but not to replace our critical thinking skills and creativity. Finding a balance in the use of AI is hard, but to me, AI is just a supplement, not a substitute for everything.

Also, when considering the ethical use of AI, it is a very sensitive opinion relating to originality, authorship, and fairness. In Tang et al Transparency in academic writing (2023), they consider authorship in academic work as attribution to humans, not AI. This means AI cannot be accountable for the integrity of the content. Nowadays, there are already a few guidelines on how to cite AI in any academic work. However, Tang et al. argue that merely citing AI is not enough if transparency about its role is not made explicit.

References

Tang et al Transparency in academic writing (2023)

From what we learned in class on Tuesday, Arrigada defines creative as novel, valuable, and surprising, a bit differently from what I think. To me, I like the definition from Jeff Goins more as it defines creativity as an artist, leader, and individual:

Artist: Seeing the world differently from others.

Individual: Creativity is unique, something that doesn’t quite fit into any box.

Leader: Creativity is the leader. It should influence others to follow a trend or a movement, shaping perceptions and inspiring innovation.

Even though these definitions are different from what we had in our readings, they still relate to each other.

Today, we had an experiment with prompting using different types of prompts and models. I asked Claude AI and Deepseek to “generate a happy poem” without any further explanation, and here are the results I received:

From what I can see, the two models generate different tones:

Claude.AI has a more energetic and fantasy tone. It is more capable of giving you a vivid and playful imagination. However, I feel like this generated poem from Claude is off-topic since it is more like an adventure poem than a happy one.

On the other hand, Deepseek generates a happy poem with a completely different style. The poem leans toward the realistic and gentle style. It focuses on everyday joys — laughter, kindness, sunlight — that readers can connect with. I feel like this version is more relevant to the topic. However, if you looking for some strike or an explosive feeling, then this poem is not for you.

Overall, with a simple-looking prompt like this, it seems like both of the models generate in a very medium level. I can’t tell which one is better because both have a different approach style. But there is one thing for sure that if my prompt had been better with more details, the models could have generated a poem that matched my expectations.

References:

A few weeks ago, I created one of my own projects, which focuses on enhancing AI use for PDF files. Basically, I used RAG ( Retrieval-Augmented Generation) to make the LLM using the data only from the imported PDF files as pieces of information for your question. Because of how general it is, it is really hard to generate a specific prompt for this problem. I spent almost a whole month and finally had my choice of prompt. Below is the prompt that I used:

“””You are an AI assistant tasked with providing detailed answers based solely on the given context. Your goal is to analyze the information provided and formulate a comprehensive, well-structured response to the question.

Because my AI does not generate answers based on a specific topic or field of study, I figured I could extract data from the PDF file as chunks of information and then store it into my data retrieval. To add more depth to my answer generation, I provided a list of questions in my prompt for the LLM so that it could also auto-generate more chunks of data from the previous data. To prevent hallucination, I told my LLM not to involve any external source: “Base your entire response solely on the information provided in the context. Do not include any external knowledge or assumptions not present in the given text.”

To ensure that the LLM knows the path of instructions to follow, I added in some of my specific commands so that the generated output would stick to what I wanted:
“””

Use clear, concise language.

Organize your answer into paragraphs for readability.

Use bullet points or numbered lists where appropriate to break down complex information.

If relevant, include any headings or subheadings to structure your response.

Ensure proper grammar, punctuation, and spelling throughout your answer.

“””

This refers to the “Cognitive Verified Pattern” which is mentioned on page 10 of the White et al. article. This ensures that the LLm would follow some of the specific requirements from the user.

AI has been evolving at an incredible pace, but the recent release of ChatGPT-4.5 feels like a significant shift. While previous iterations brought major improvements, GPT-4.5 introduces refinements that make AI interactions feel more fluid, reliable, and, dare I say, almost human.

So, what makes ChatGPT-4.5 stand out?

One of the biggest changes is its enhanced contextual understanding. GPT-4.5 processes longer conversations with better memory, meaning it can recall details more accurately throughout a discussion. In contrast, older versions struggled with maintaining consistency, often forgetting information or repeating answers unnecessarily.

Another notable upgrade is the reduction of hallucinations—when AI generates misleading or entirely false information. While GPT-4 still had moments of confident inaccuracy, GPT-4.5 shows marked improvement in factual accuracy, thanks to better training data filtering and reinforcement learning.

Moreover, the multilingual capabilities have been refined. While GPT-4 supported multiple languages, GPT-4.5 delivers more natural translations and culturally aware responses. This makes it an even better tool for global users.

Performance-wise, GPT-4.5 is also faster and more efficient, even with its increased complexity. It generates responses quicker and requires less computation per query, making interactions smoother.

However, it’s not just about speed and accuracy—the conversational tone is more dynamic. GPT-4.5 adapts better to emotions and writing styles, making responses feel more natural and engaging.

So, is GPT-4.5 a game-changer? It’s definitely a step forward, though we’re still far from a perfect AI. But one thing is clear: the gap between human and AI interactions is shrinking faster than ever.

Feel free to check out news about GPT-4.5 at https://www.wired.com/story/openai-gpt-45

Hi, my name is Quoc Do (he/him), from the class of 2026,  and I’m a Computer Science major. I enjoy playing games, cooking, going to the gym, and coding, especially when exploring AI applications. My Independent Study (IS) research is deeply related to AI, so I’m taking this class to strengthen my knowledge in this area. I’m looking forward to learning and collaborating with everyone!

AI is a big part of our lives today, helping with things like recommendations, virtual assistants, and automation. It makes tasks easier and improves decision-making in areas like healthcare and finance. However, as AI grows, we also need to think about important issues like privacy, bias, and job changes to make sure it’s used effectively.