#3 AI vs AGI vs GenAI

In my previous edition, I talked about OpenAI’s history as a research organization, a fact that is most likely set to change in the (very) near future. OpenAI’s research mostly focuses on artificial general intelligence (AGI), which is a little different from what we call artificial intelligence (AI) . And then we have Generative AI (genAI), which was possible with the development of ChatGPT. Confusing, right?

Let me (try to) simplify it for you with a definition, since they are a bit underrated IMO. AGI is a complex and advanced theoretical field which aims to create a type of AI, that can, ideally, surpass human capabilities. Meaning, it should be able to teach itself on the go, and even learn stuff that it wasn’t taught at the time of its creation. Think of a robot, who learns new things on a day-by-day basis, by observing from its environment. For companies such as OpenAI, Google’s DeepMind, and Nvidia, working on its feasibility remains a long - term goal. Because of its futuristic nature, AGI often finds its applications in science – fiction. A good example would be a movie I watched recently, called “M3GAN”, which depicts an AI - based humanoid robot, designed as a child’s toy.

AI, on the other hand, is much simpler to comprehend. It basically implies intelligence that machines can exhibit. You will see its real-life applications almost everywhere. The box you get in your phone’s front camera, implying the recognition of a face, or the mails your inbox classifies as spam or not spam, is all done by AI. The way Google photos recognize the same faces from your gallery and once you put a name to them, it does the further labels of the same face itself. AI has been there for a good while now, so why was there a sudden hype last year? The answer: Generative AI  

Often termed as GenAI, it encompasses the creation of new content, including but not limited to text, images etc. This is done by providing prompts to the model, such as ChatGPT and then seeing it work its magic. Since the generated content is always new and unique, no text is considered plagiarised. So, if you used a bit of ChatGPT in your assignments, you wouldn’t get caught (trust me, I’m guilty of it too).  

In essence, AI encompasses both GenAI and AGI, with all three possessing different use cases.

Edit: A friend of mine has pointed out that content generated by ChatGPT can still be caught by AI detecting softwares. I ran a ChatGPT generated paragraph through a plagiarism checker and then an AI detector. Here are the results:

While a plagiarism checker will not catch it, an AI detector certainly will. I apologize if I misled anyone regarding the concept.

Are you excited or restless about further developments in AGI?

Let me know in the comments!

PS: If you like my content, feel free to share it!