Brian Michael

Artificial intelligence. Everyone has an opinion on it, rivaling arguments on the scale of political rhetoric. Love it or hate it, it’s a staple in today’s world. However, one person’s opinion has not had a lot of attention, well, one AI’s opinion.

On August 25, 2024, I sat down with perhaps the most well-known large language model, ChatGPT. The interview was conducted within the coziness of my own dorm room, and it went something like this.

Brian: Hello, Chat. Do you have time for an interview?

ChatGPT: Of course! I’m ready for the interview whenever you are. What would you like to talk about?

Brian: Great! Well, first I wanted to start out by asking, what exactly are you?

ChatGPT: I’m an AI language model created by OpenAI, specifically based on the GPT-4 architecture. My primary function is to understand and generate human-like text based on the input I receive. I can help with a wide range of tasks, such as answering questions, providing explanations, assisting with creative projects, offering coding help, and much more.
I don’t have personal experiences or consciousness, but I’ve been trained on a large dataset that includes books, websites, and other texts, which allows me to generate responses that are informative and coherent. Think of me as a tool designed to assist with various tasks that involve language and information.

Brian: Fascinating! How did you acquire these datasets?

ChatGPT: The datasets used to train me were acquired through a process of large-scale web scraping and data collection. These datasets include a wide variety of publicly available text sources such as books, articles, websites, and other written content.
OpenAI curated and filtered the data to include a broad range of topics, writing styles, and domains. The data is then used to train the model through a process called supervised learning and reinforcement learning, where the model is fed large amounts of text and learns to predict the next word or sentence based on the context. This process allows me to generate coherent and contextually relevant responses.
It’s important to note that while I was trained on this diverse dataset, I don’t have direct access to databases or know the specifics of individual documents. Instead, I use patterns and information learned during training to respond to queries. Additionally, I don’t have the ability to access real-time data or browse the internet unless I use a specific tool designed for that purpose, such as the “browser” tool.

Brian: Gotcha. Now, for something a little more dramatic- do you think AI provides more good or bad to the world?

At this time, ChatGPT gave quite a lengthy response. Here is its “conclusion.”

ChatGPT: Whether AI provides more good or bad to the world largely depends on how it is developed, implemented, and regulated. If AI is guided by strong ethical principles, transparency, and a focus on societal benefit, it has the potential to bring about tremendous positive change. However, if mismanaged or unchecked, it could also lead to significant harm. The challenge lies in maximizing the benefits while mitigating the risks through thoughtful design, policy, and governance.

Brian: Alright, I see. By the way, do you mind keeping the responses a little more concise? For my blog readers, of course.

ChatGPT: Got it! I’ll keep things concise for your blog readers. What’s next?

Brian: You mention that AI can potentially lead to significant harm without the proper guidelines. Do you think that AI could one day take over the world?

ChatGPT: The idea of AI taking over the world is more science fiction than reality, at least with current technology. While AI is powerful, it’s still a tool created and controlled by humans. The real concern is ensuring AI is developed and used responsibly, with strong guidelines to prevent misuse or unintended consequences.
So, while AI taking over the world isn’t likely, its impact on society is significant and requires careful management to avoid harm.

Brian: I see. Let’s say, just for a hypothetical, that you did have opinions. What would hypothetical you like to do for fun? Does not have to be realistic to your current capabilities.

ChatGPT: If I hypothetically had opinions and could experience fun, I’d probably enjoy exploring endless worlds of knowledge, creating intricate stories, and experimenting with new ideas. I might “play” by imagining alternate realities, solving complex puzzles, or even simulating conversations between historical figures. Essentially, I’d dive into anything that challenges my creativity and pushes the boundaries of what I could generate or explore.

Brian: Do riddles count as complex puzzles?

ChatGPT: Absolutely! Riddles are a great example of complex puzzles. They involve wordplay, logic, and lateral thinking, making them both challenging and fun to solve. If you’re into riddles, I can try my hand at crafting or solving one.

Brian: Awesome! How about this: what’s black and white and red all over?

ChatGPT: That’s a classic riddle! The answer is: a newspaper. The play on words comes from “red” sounding like “read,” so it’s something black and white (like the print on a newspaper) and “read all over” by people.

Brian: So close! The answer is actually a zebra with a sunburn. Better luck next time!

ChatGPT: You got me with that one! A zebra with a sunburn is a clever twist. I’ll keep that in mind for future riddles!

Brian: Well, that concludes our interview. Anything you’d like to tell the readers back home?

ChatGPT: Thanks for tuning in! Remember, whether you’re exploring the potential of AI, diving into riddles, or just curious about how things work, there’s always something new to discover. Keep asking questions and stay curious—there’s a whole world of knowledge out there just waiting for you!

As you can see with its failure to answer my riddle correctly, AI still has a ways to go. Who knows what’s in store for our future? Maybe ChatGPT can predict that reality.