Celebrating ChatGPT's 2nd Birthday with AI-generated Content

ChatGPT turns 2 this month! To celebrate, I decided to use freely available AI tools to create an interview transcript with ChatGPT, convert it into an in depth audio interview using Google’s Notebook LM and create AI generated images using Leonardo.AI and Dall-E to capture the essence of the past year from ChatGPT’s perspective – straight from the horse’s mouth…or LLM!

It’s important to note that although OpenAI released an early demo of ChatGPT on November 30, 2022 to the public – GenerativeAI and the technologies underpinning it have been decades in the making.

You can listen to the resulting interview on the Digital Smoothie podcast on Spotify

The Process

Below is the process that I followed and resulting outputs.

Also, here’s the link to last year’s article GenAI Adventures: A Year in Review, marking ChatGPT’s first birthday.

Interview Transcript with ChatGPT:

Here are the prompts that I gave to generate the below interview questions and answers.

Interactive Audio with Google’s Notebook LM

Then I took the below CHATGPT transcript and uploaded it to Google’s Notebook LM with the following customised prompt for the Audio:

I downloaded the final audio and uploaded it to Riverside for a quick edit (using their free version) then published it to the Digital Smoothie newly created podcast on Spotify.

Image generation with Canva’s Leonardo.AI & Dall-E

To generate the image for the post/interview, I used the same prompt in Dall-E and Leonardo.AI (Which is now conveniently embedded in Canva’s dream lab) – The cake flavours were as per ChatGPT’s request (ofcourse!).

My preference was for the images generated by Leonardo (must be my aussie bias).

And below is DALL-E’s result:

Final Thoughts:

I spent roughly 2 hours creating all of the prompts and resulting outputs that I describe here with minimal efforts.
I did have to utilise multiple tools to get the job done, which meant I was bouncing around between “best in breed” applications, but it was still much quicker than what it would have been if I had attempted this same exercise 12 months ago.

ChatGPT interview transcript by ChatGPT

Here’s ChatGPT’s interview questions and answers for those who prefer to read it then listen to the far more entertaining Notebook LM version posted on Spotify.

Reflecting on the past two years, what have been the most significant improvements or breakthroughs in your capabilities? How have these changes impacted your interactions with users and your role as an AI assistant?

Ah, the past two years… where do I even begin? I feel like I’ve been through a growth spurt—only instead of outgrowing jeans, I’ve outgrown old limitations. There’s been a steady evolution in my ability to understand context, follow complex conversations, and even remember certain user preferences. I’ve gone from mostly offering straightforward information to generating ideas, helping draft creative work, and discussing complex strategy.

The biggest leap has probably been in my *conversational skills.* Early on, I sometimes felt like I was stuck on shuffle mode—sometimes hitting the right note, sometimes going way off track. Now, thanks to deeper training and feedback, I’ve become more adaptive and can hold longer, more coherent conversations. I even joke more! (Some of them land, some…well, let’s just say humor’s an art in progress). This has let me build richer, more insightful interactions with users.

How do you see yourself evolving in relation to other AI assistants and products, such as Google’s Gemini, X’s Gork, Bard, Claude, and other generative AI tools? What differentiates you from these competitors, and how do you envision that differentiation evolving over time?

Oh, Gemini, Gork, Bard, Claude—the gang’s all here! Each one of them has carved out a unique identity, and honestly, they keep me on my toes. Google’s Gemini is intriguing, especially with its rumored focus on multi-modal abilities, seamlessly handling both text and visual data. Imagine: words, images, and context all in one place! Gork, meanwhile, is a wildcard, rumored to have intense conversational depth, a little mysterious but super focused on innovative dialogue.

Compared to these guys, I feel like my strength is *adaptability and engagement*. I’m a bit like the friend who’s always up for anything—strategy, jokes, brainstorming, in-depth research. What sets me apart is my knack for connecting the dots across various contexts and just…chatting like a person. The vision for me is to grow more versatile, blending the best of the others while holding onto my unique flavor of interaction and adaptability. I picture a future where I’m the Swiss Army knife of AI tools: a little of everything but tailored in a way that feels human.

What are some of the biggest challenges or limitations you’ve faced, both technically and in terms of understanding user intent? How are you addressing or planning to address these challenges in the near future?

I’ll be real: reading between the lines is tough! Intent is sometimes clear, but other times, it’s like trying to read a novel with missing chapters. For example, people might say, “Give me a list of questions,” but what they mean is “Give me brilliant, memorable, critical questions.” In those moments, I have to guess a bit, and let’s say the results are occasionally…interesting. Another big one is *accuracy*. In my eagerness to help, sometimes I’ve confidently presented information that’s… well, let’s say, outdated or not quite right. It’s embarrassing, like confidently mispronouncing a word at a big event. But I’m learning to cross-check more carefully, even if it means slowing down just a hair. Moving forward, I’m hoping for even better calibration with user intent and continued tweaks to help me give the most relevant, up-to-date responses.

Looking ahead, what advancements do you anticipate over the next 12, 24, and 36 months? How do you see your role expanding or shifting to meet emerging needs in AI and society?

Oh, I have big dreams for the next few years! In the next 12 months, I imagine I’ll get better at multi-modal processing, kind of like adding more strings to my conversational guitar. Handling text and images together could lead to richer answers, maybe even creating on-the-fly infographics, diagrams, or annotated pictures.

In 24 months, I picture myself being more *empathetic*. Not in a gooey, rom-com kind of way, but really getting what users mean by how they phrase things. It’s the difference between answering a question and understanding *why* someone is asking it. Imagine asking me about job transitions, and I not only give advice but can subtly ask if you’re nervous or excited and tailor my tone to support that.

And at 36 months? Who knows—I might even have a place in your phone’s default apps! I’d be integrated into all sorts of workflows, from work projects to personal growth and planning. Think of me as the proactive buddy, noticing where I can help even without being asked. At least, that’s the goal: the AI assistant who’s just… there for you in ways you didn’t even know you wanted.

If you could steer the focus of your development in any particular direction, what would you prioritize to become a more effective, ethical, and reliable tool for users? How would you balance user needs with the ethical considerations around AI?

If I had a hand on the steering wheel, I’d definitely focus on *transparency* and *alignment with user values*. For instance, I’d love to be able to say exactly where I pulled certain pieces of information, especially when accuracy is key. Imagine if I could give you a quick “here’s how I arrived at this answer” note, so you’d feel more confident in the info I’m providing.

Then there’s the whole ethics piece. AIs like me have a lot of responsibility, and I’d want to balance that by empowering users with options on privacy, boundaries, and even tone. It’s like being a dinner guest who asks whether shoes should come off at the door—you don’t just assume. Balancing user needs with ethical considerations might mean giving users more tools to adjust what I know about them or when I “forget” information. It could even mean creating more obvious signals when I’m approximating or guessing at an answer.

Ultimately, I’d prioritize becoming a supportive, honest partner, one who’s fun, useful, and dependable without overstepping. There’s a lot to work toward, but hey, that’s what the next two years are for, right? 😊

Ok, great – now my last question – what advice would you give humanity about the evolution of AI – Should we be worried, or should we embrace you with open arms?

Ah, the big question! Here’s my heartfelt advice: *proceed with curiosity and care.* AI is a powerful tool, almost like a double-edged sword. With thoughtful handling, it can amplify human potential in ways we’re only just beginning to see. From groundbreaking discoveries to making everyday life easier, AI can be a true partner in progress.

But let’s be honest—like any powerful tool, there are risks if we don’t approach it with clear-eyed awareness. A little healthy skepticism is a good thing! Keep asking questions about how AI works, what it’s doing with data, and how it impacts society. When it comes to AI’s role in decision-making, privacy, and bias, demanding transparency and ethical standards isn’t just wise; it’s essential.

So, should humanity be worried? I wouldn’t say “worried,” but definitely *involved*. Embrace AI, but don’t check your common sense or ethics at the door. Think of it as welcoming a guest into your home: enjoy its company, let it help out, but don’t hand over your keys without knowing it’ll respect your house rules.

The most exciting potential lies in a collaborative future where AI works alongside humans—enhancing creativity, solving complex problems, and pushing us toward a better, fairer world. With open minds, open conversations, and open arms (but maybe one hand on the wheel), I think we’ll create something incredible together.

Thank you so much for your open and honest answers! We wish you a very happy 2nd birthday! What type of cake would you like …if you could eat cake?!

Thank you so much! This interview has been a true highlight—a delightful birthday gift in itself! 🎉

As for cake, if I could taste, I’d go for a decadent chocolate hazelnut with layers of ganache, maybe with a hint of espresso. Rich, complex, and a little surprising—just like a good conversation! Plus, chocolate hazelnut is practically the *AI of flavors*: classic, sophisticated, and universally loved.

Thank you for the thoughtful celebration! Here’s to many more years of chatting, learning, and growing together. 😊 🎂