What are ChatGPT and Generative AI (and how can I use them)?

There’s been no shortage of articles on ChatGPT since it was launched in November 2022 (so apologies for adding one more on the stack). From CNET to CNBC, everyone is talking about the new “AI arms race” that ChatGPT has sparked.

But even with all this coverage, not a lot of people still know the specifics of the technology. What is ChatGPT and this ‘Generative AI’ people talk about? Is it going to replace Google? Is it factually accurate? And perhaps the big one, is this the building blocks of general intelligence — the name for an AI that has intelligence comparable to a human?

In this article, we’ve gathered up the most frequently asked questions about ChatGPT on Google and Reddit to answer them all, no matter how silly they seem.

Generative Artificial Intelligence is any type of AI that can be used to create new and original content based on patterns and examples it has learned. This content can be text, images, video, code, or synthetic data. Examples include DALL-E, Midjourney, and ChatGPT.

An example is an AI trained on trees might be able to create images of trees that don’t exist in the real world, based on patterns it has learned. These AIs often take written prompts from humans as input, and turn it into the desired output. Generative AI models are often used in unsupervised machine learning problems.

Discriminative AI is a different type of AI that is trained to recognise and classify patterns in existing data. Like its namesake, it “discriminates” between classes and categories based on examples it’s learned. 

An example would be choosing if a picture has an image of a cat or a dog. Generative AI has been used for things like fraud detection (“Is this person engaging in sketchy behavior?”) to image recognition. Discriminative AI is normally used for supervised machine learning.

Image Source: betterprogramming.pub

ChatGPT stands for “Chat Generative Pre-Trained Transformer”, and it’s a generative AI language model that acts in a conversational way. You can ask it questions and get human-like answers. It is developed by OpenAI.

You can use it to do all sort of things, such as:

• Explain calculus in a simple, easy to understand way

• Write a paragraph explaining the history of computer science using a friendly tone

• Optimize sections of your code, or generate entirely new code

• Write a haiku about pudding

• Suggest baby names

Really, the sky's the limit! It is, in essence, a super-powered artificial intelligence chat bot.

Yes, it can write Haiku. Is it any good at it? That’s an entirely different story…

ChatGPT has been trained on a massive dataset which includes the whole of Wikipedia, scholarly articles, research papers, news articles, books, and documentation.

It is important to note that ChatGPT doesn’t do an internet search for answers; it predicts answers based on its training data! 

ChatGPT is also limited to data that was available at the time it was trained. If you ask it about something that happened this month, for example, it may have no idea what you’re talking about (or give you an outdated answer).

At least ChatGPT is honest if you ask it upfront.

Yes, as of March 2023, ChatGPT is free to use. There is a pilot subscription plan for ChatGPT called ChatGPT Plus, which is available for $20/month. Subscribers gain constant access to ChatGPT even during peak times, faster response times, and priority access to features and improvements.

ChatGPT will likely be using a freemium model for the immediate future. OpenAI has openly stated it will be “actively exploring options for lower-cost plans, business plans, and data packs for more availability.” Whether or not it continues offering a free version remains to be determined.

Given that ChatGPT uses a Large Language Model that must be trained on available text materials, which includes Wikipedia, it arguably can’t replace what it is sourcing from. However, there is debate amongst Wikipedia authors about whether or not ChatGPT should be able to write Wikipedia articles, or whether Wikipedia should be a “human-only” zone.

While there is a lot of hype about ChatGPT being a “Google Killer,” since ChatGPT does not do an internet search to provide information, it cannot currently replace Google. Pundits are quick to point out that while ChatGPT can be used to answer questions like Google, it was not its intended design. That said, Google has certainly been spooked by the rise of ChatGPT, and it is working on its rival system called Bard — even though it has had a few hiccups.

While ChatGPT can give correct answers, it can also give “confidently wrong” answers. Answers by ChatGPT should not be assumed to be automatically accurate, so it is important to leverage your own critical thinking and research skills to confirm what it is saying is true. 

For evidence of this, we strongly recommend watching this video by Simon Allardice, where he demonstrates how ChatGPT can get things confidently wrong (and then stubbornly stick to its guns).

... Also, ChatGPT is pretty sure you shouldn’t rely just on ChatGPT (Or is this the Liar Paradox?)

While there is a general consensus among experts that ChatGPT is not an example of human-level intelligence (also known as “artificial general intelligence” or “strong AI” — it doesn’t quite fit into the definition of traditional AI products (“narrow artificial intelligence” or “weak AI.” Instead, it falls into a murky category between them.

“(ChatGPT) shows that pouring more data and more computing time and power into the deep learning paradigm can lead to astonishing results. The fact that GPT-3 is even worthy of an “is this AGI?” conversation points to something important: It signals a step-change in AI development,” said Gary Grossman, Senior VP of Technology Practice at Edelman and Global Lead of the Edelman AI Center of Excellence. 

“If nothing else, GPT-3 tells us there is a middle ground between narrow and general AI. It is my belief that GPT-3 does not perfectly fit the definition of either narrow AI or general AI. Instead, it shows that we have advanced into a twilight zone. Thus, GPT-3 is an example of what I am calling ‘transitional AI.’”

According to a survey of industry experts by Business Insider, the following jobs that were most likely to be impacted by ChatGPT and similar future generative AI were:

• Technology jobs (Coders, computer programmers, software engineers, and data analysts)

• Media jobs (advertising, content creation, technical writing, journalism)

• Legal industry jobs (paralegals, legal assistants)

• Market research analysts

• Teachers

• Finance jobs (Financial analysts, personal finance advisors)

• Investment traders

• Graphic designers

• Accountants

• Customer service agents

As you can see, these make up a large chunk of the job market! However, a silver lining is that these technologies are forecasted to not replace but assist these roles — programmers everywhere are already starting to consult with ChatGPT for ways to optimize their code, and authors are using ChatGPT for brainstorming purposes.

ChatGPT is currently free, and you can check it out — playing around with it is a great way to get started. However, if you want a more structured and less random approach to learning about what you can do with ChatGPT, we recommend checking out Amber Israelsen’s free course: “ChatGPT and Generative AI: The Big Picture.” 

If you’re a programmer who’d love to learn about how to work with ChatGPT, our resident guru Jeremy Morgan has also done a short video on how to do just that. We will also be covering this topic more in depth in an upcoming article.