I “Interviewed” ChatGPT And Here’s What It Has To Say

Artificial intelligence is a term that has been misused over the past decade, IMHO. The words have been thrown around a lot with consumer electronics that, upon using the gadgets, it make you wonder “where’s the AI?” Then last year, OpenAI founded by a bunch of dudes in 2015, including maverick entrepreneur Elon Musk, launches ChatGPT and kind of put AI into overdrive.

GPT in ChatGPT stands for Generative Pre-trained Transformer. It is a chatbot launched in November 2022. Instead of telling you that is built on top of OpenAI’s GPT-3 family of large language models and is fine-tuned with both supervised and reinforcement learning techniques, we will let you hear it from the horse’s mouth in an “interview” I had with it. Just kidding. It can’t do an interview but it can surely answer questions and do it very well.

Can you explain, in simple terms, what is ChatGPT?

“ChatGPT is a type of artificial intelligence program that is able to understand and respond to natural language text. It uses a type of machine learning called deep learning to analyze and understand the meaning of the text it receives and then generates its own text in response. It can be used for a variety of purposes, such as chatbots, automated customer service, and text generation.”

In short, it is the next level of AI that you can have a proper dialog. And it does it way better than chatbots from certain telcos. Expect to get less of “Sorry, I do not understand your question. Can you rephrase it?” (sounds familiar?). And it will respond in a less, well, robotic manner. It was specifically trained to interact in a conversational way which means it is possible for it to answer follow-up questions. And if it was wrong, it will readily admit it, though I have yet to experience it yet myself. In greater depth, here’s what this AI-powered chatbot is about, according to it:

Can you explain further?

“At a high level, ChatGPT is a machine learning model that is trained on large amounts of text data. It uses a type of neural network called a transformer to analyze the text and understand the relationships between the words and phrases in it.

Once the model has been trained, it can be used to generate text that is similar to the text it was trained on. For example, if the model was trained on a dataset of customer service conversations, it can generate responses that sound like they were written by a customer service representative.

It can also be used to complete a text given a prompt. For instance, if you give it a sentence as a starting point, it will continue writing a coherent and relevant text based on that.

Overall, ChatGPT is a powerful tool for understanding and generating natural language text, and it has a wide range of potential applications in industries such as customer service, marketing, and content generation.”

For tech buffs and futurists, you may have picked up the power ChatGPT at this point. It needs not a regurgitation of what you have asked. It can pick up from the last question and work on it to give me an answer. It was quite a mind-blowing experience.

I have conversed it with several times. Personally, I find it yet not as convincing but definitely way better than most chatbots you find today’s businesses are using. Though I have to admit, I sometimes forgot that I was talking to a machine. Somehow, at some point, I find myself imagining William Daniels’ voice of the original Knight Rider K.I.T.T. responding to my questions. And on the question of if it, or AI in general, will take over the world, here’s what it has to say:

Do you think, ChatGPT, and AI in general, will take over the world?

“AI has the potential to greatly improve many aspects of society, but it is unlikely that it will “take over the world” in the sense of becoming an all-powerful entity that controls everything.

While AI systems like ChatGPT can be very advanced and perform tasks that were previously only possible for humans to do, they are still limited by the data they were trained on and the specific tasks they were designed to perform. They do not have consciousness, emotions or the ability to make moral or ethical decisions.

Additionally, there are many people working to ensure that AI is developed and used in a responsible and ethical way, and there are laws and regulations in place to govern its use.

It’s important to note that AI systems are ultimately tools created and controlled by humans, and it’s up to us to ensure that they are used for the betterment of society, rather than causing harm.”

I can totally imagine Elon Musk saying the last part. After all, the man who help started OpenAI also warned us about AI. Anyhoo, at some point in the “interview”, things can become rather very robotic. When prompted about what it can do, here’s the answer:

But what can ChatGPT do?

“ChatGPT is a language model, which means it is able to understand and generate natural language text. Some specific tasks that ChatGPT can be used for include:

Text completion: Given a starting sentence, ChatGPT can generate text that continues the sentence and makes it coherent and relevant text.
Text generation: Given a prompt, ChatGPT can generate new text that is similar to the text it was trained on.
Chatbot: By training ChatGPT on a dataset of conversations, it can be used to generate responses to customer service inquiries or other types of interactions.
Language Translation: By fine-tuning the model on a specific language pair, it can be used to translate text from one language to another.
Summarization: By fine-tuning the model on a specific task of summarizing, it can be used to reduce the length of a text by retaining its main points.
Text classification: By fine-tuning the model on a specific task of classifying text, it can be used to categorize a text according to certain criteria.

This list is not exhaustive, but it gives an idea of the types of tasks that ChatGPT can perform. Because it is a general-purpose language model, it can be fine-tuned and adapted to a wide range of natural language processing tasks.”

Tell me that is not a very robotic answer. But then, this answer to my next question brings back the K.I.T.T. of my imagination again. I asked ChatGPT its thoughts on Microsoft Bing’s plan on integrating ChatGPT into its search engine:

What’s your thought on Microsoft Bing integrating ChatGPT into its search engine?

“As an AI model, I don’t have personal thoughts, but I can tell you that integrating a language model like ChatGPT into a search engine like Bing has the potential to greatly improve the user experience.

By incorporating a language model, Bing could understand and respond to more natural, conversational queries, making it easier for users to find the information they’re looking for. Additionally, by generating text that is similar to the text it was trained on, ChatGPT could be used to provide more detailed and accurate answers to users’ questions.

It could also be used to generate more human-like summaries of the search results, making it easier for users to understand the main points of a webpage without having to read through the entire thing.

However, it’s worth noting that as with any technology, it’s important to ensure that the integration is done in a responsible and ethical way, with proper privacy and bias considerations in place.”

Now, that is a very human-like answer, IMHO. Anyhoo, in case anyone’s wondering, no. It is not the peak of AI. As a reminder, it is a language model, and therefore, it can only respond to queries and not initiate a conversation without human prompting. But man, the way it responds is so human-like. And finally, we asked about the pricing:

“OpenAI is a research organization and its business model is to make its AI research and models available to the public while also generating revenue. The exact pricing and plans for using specific models like ChatGPT have not been announced yet.

Currently, OpenAI provides access to ChatGPT via its API, which allows developers to use the model to build applications and services. OpenAI also offers a free tier for developers to try out the model, and a paid version for more heavy usage.

Please note that this information is based on my knowledge cutoff and that OpenAI’s pricing and plans are subject to change. I recommend checking OpenAI’s website for the most up-to-date information on their pricing for using ChatGPT.”

I find that answers to a question that leans toward business queries are more robotic. Anyhoo, this means expect chatbots of businesses will get better in the future. Hopefully less of the “I am sorry, I do not understand the question. Can you rephrase it?”

ChatGPT has a significant impact on blogging too. In the past, we do not need to worry about AI-generated articles but now, there is a chance that some articles are written by AI. Speaking of which, we heard, Google is not liking it. I asked ChatGPT if it could write an article for me, and it promptly replied with ‘yes’. It can help me write an article on a topic of my choice but the article is limited to the knowledge it learned from data that it was trained on.

One final question I have for ChatGPT:

I mean, would you be able to get onto the Internet and start to learn things on your own?

As a language model, I do not have the ability to access the internet and learn on my own. I can only provide responses based on the patterns and relationships I learned from the data that was used to train me. The training data is a large corpus of text data that was available up to my knowledge cut-off date, which is 2021.

It’s important to note that my abilities are determined by the architecture and parameters of the model, as well as the quality and quantity of the data used to train it. While some AI models may have been designed to learn from the internet, ChatGPT is not one of them.

Additionally, there are also ethical and security concerns with AI models having internet access, such as privacy, bias, and malicious use, which is why it’s important to have proper governance in place.

Please let me know if there’s anything else I can help you with.

So, there’s no worry about AI. At least not until some mad scientist found a way to give it consciousness. If that happens, be prepared to meet the demand for AI and/or robot rights. I mean, that’s what we thinking people love to ask for right?

Featured photo by Antony Trivet on Pexels. Edited by Mikeshouts.