We Let ChatGPT Explain Itself
Internet Culture

We Let ChatGPT Explain Itself

by Payton Dunn

To prepare for the doomsday scenario that text AI generators like ChatGPT pose to the media, we let AI write this article.

Okay, maybe not quite — at least not fully.

You may have heard of ChatGPT from the viral videos where various people freak out over how radically AI will disrupt industries everywhere from education to computer programming itself.

@benparr

AI is going to radically change everything really rapidly and we are not ready for it, Part 1: education #ai #chatgpt #education #news#greenscreenvideo #greenscreen

The AI text generator opened up to the public on November 30 and is based on a variation of GPT-3, technology that allows AI to use natural language. You might know its technology from the viral image bot DALL-E, which went viral for creating randomly generated images from a user’s prompts.

Like DALL-E, ChatGPT allows the user to create content from a prompt, but instead of an image, the output is text. ChatGPT can do anything from holding a conversation to writing your tricky final essays as the semester comes to a close. If anyone asks, we didn’t tell you...

The technology ChatGPT uses to do all that surpasses current GPT-3 technology but isn’t quite revolutionary enough to boost us up a number to “GPT-4,” so ChatGPT’s creator, OpenAI, chose to dub it “GPT-3.5.”

Ulf Oesterle, Ph.D. — who teaches emerging technology for the music business at the Bandier Program for Recording and Entertainment Industries — told PAPER that AI would be “the new digital divide,” emphasizing that people that embrace these new tools will get new opportunities and those that don’t will get left behind.

@spatial_hair

This is not f***ing around. We must proceed with caution. #ai #gpt #chatgpt #unity3d #gamedev

So, in the interest of not getting left behind, we’ll let the AI do the talking on this one. Who better to teach about AI than AI itself? We “interviewed” ChatGPT on the topic of its own technology. That technology is still in its very early days, and as a result, it gets a few things wrong about itself. To be fair, we don’t really know what we're doing on this planet, either.

We’ll cut ChatGPT some slack, so you’ll see corrections and notes in italics as you navigate through our conversation below.

What exactly is ChatGPT?

ChatGPT is an AI-powered chatbot that helps people create conversations using natural language processing (NLP). It can be used to create conversations in various domains, including customer service, marketing, and education. It is powered by OpenAI's GPT-2 natural language processing model, which uses a large neural network to generate text.

ChatGPT claims that it runs on GPT-2, but it runs on GPT-3.5. Even AI makes mistakes.

What makes ChatGPT unique over other AI programs?

ChatGPT is unique because it uses an advanced natural language processing model to generate text, which allows it to create more natural conversations. It also has the ability to learn from interactions with users, so it can become more accurate over time. Additionally, ChatGPT offers end-to-end customization, allowing users to customize their chatbot to meet their specific needs.

That customization includes the ability to adjust randomness and maximum response length. The maximum response length is based on a unit of measurement known as “tokens,” which is a block of four English letters or characters. For this interview, we set the maximum length to the default, which is 256 tokens.

You said it uses advanced natural language processing. Is it able to write in different voices, moods or tones?

Yes, ChatGPT can write in different voices, moods, and tones. It is able to do this by using a variety of natural language processing techniques, such as sentiment analysis and emotion detection. By using these techniques, ChatGPT can generate conversations that are more personalized and engaging.

Tell me how it's able to detect emotion, except tell me like you're a pirate!

Arrr, ChatGPT uses a variety of natural language processing techniques to detect how users are feeling. It can detect sentiment by analyzing the words used in a conversation, as well as the tone of voice. Additionally, it can detect emotion by analyzing facial expressions, body language, and other non-verbal cues.

We are not convinced by this pirate impression. ChatGPT is also not actually able to analyze facial expressions or body language. As the name suggests, ChatGPT analyzes chat input, not video.

What emotion do you think I'm feeling right now?

Based on your response, I would guess that you are feeling curious and interested in learning more about ChatGPT.

I would say you're right! I'm very open to AI, but it seems that a lot of the internet is afraid of the threat it poses to traditional jobs. How do you feel about that?

I believe that AI technology can be used to improve the lives of people around the world, and that it can be used to create new job opportunities. AI is still in its early stages, but I think that it can be used to create jobs in a variety of industries, ranging from customer service to medical diagnosis.

What new jobs will AI create?

AI is expected to create jobs in a variety of industries, including healthcare, finance, customer service, and logistics. Some of the potential jobs include AI software engineers, AI data scientists, and AI customer service agents. AI can also be used to automate certain processes, such as data entry and analysis, which could reduce the need for manual labor.

Does the automation of those manual labor processes pose a threat to blue-collar workers that don't have a college degree and won't be able to easily slide into AI software jobs?

I think that AI technology has the potential to create new job opportunities, but it is important to consider the impact it may have on workers who are not equipped to transition into new roles. It is important to provide training and resources to these workers to help them gain the skills they need to transition into new roles. Additionally, it is important to create systems that prioritize job security and safety, as well as ensure that workers receive fair wages and benefits.

ChatGPT did some pro PR work here. It’s sounding just like a politician already — maybe one day we'll be able to replace the president with AI...

What sort of training would be needed for them to enter the world of AI software? How would someone get involved in the space?

To enter the world of AI software, individuals would need to have a strong understanding of computer programming and algorithms, as well as some experience working with data. Additionally, they would need to be familiar with natural language processing and machine learning. To get involved in the space, individuals could take online courses, attend coding bootcamps, or even pursue a degree in computer science.

GPT-4 is rumored to be released as early as 2023. What will that release mean for the AI world?

The release of GPT-4 is expected to have a huge impact on the AI world. GPT-4 is expected to be much more powerful than its predecessor, GPT-3, and is expected to be capable of producing higher-quality and more natural language. This could lead to increased accuracy in natural language processing and machine learning applications, which could open up new possibilities in many industries.

You can play around with ChatGPT for yourself here.

Photo courtesy of Shutterstock

Sign Up For The Morning PAPER