AI terminology explained: the dictionary of artificial intelligence

The field of AI is so full of jargon that it can be very difficult to understand what is really happening with each new development. And the truth is that AI is everywhere, but not everyone understands what we are talking about when we refer to it.

ChatGPT DOWNLOAD

To help you better understand what’s going on, The Verge has compiled a list of some of the most common AI terms, and we’re bringing it to you translated and adapted to your language.

First, let’s define what AI is (we talked about Agentic AI the other day) and then we will explain each term related to artificial intelligence. Let’s go!

What is AI exactly?

Often abbreviated as AI, it is technically the field of computer science dedicated to creating computer systems that can think like a human being.

But nowadays there is a lot of talk about AI as a technology and even as an entity, and it is difficult to pinpoint exactly what it means. It is also frequently used as a buzzword in marketing, which makes its definition more fluid than it should be.

Google, for example, talks a lot about how it has been investing in AI for years. This refers to how many of its products improve thanks to artificial intelligence and how the company offers tools like Gemini that seem intelligent, for example.

There are underlying AI models that power many AI tools, such as OpenAI’s GPT. Then there is Meta’s CEO, Mark Zuckerberg, who has used AI as a noun to refer to individual chatbots.

As more companies try to sell AI as the next big thing, the ways they use the term and other related nomenclature can become even more confusing.

AI Terminology

Machine Learning: Machine learning systems are trained (we will explain training later) with data so they can make predictions about new information. This way, they can ‘learn.’ Machine learning is a field within artificial intelligence and is fundamental for many AI technologies.

Artificial General Intelligence (AGI): Artificial intelligence that is as intelligent or more intelligent than a human. OpenAI, in particular, is investing heavily in AGI. It could be an incredibly powerful technology, but for many people it is also potentially the most terrifying prospect about the possibilities of AI: let’s think about all the movies we’ve seen about super-intelligent machines taking over the world! To make matters worse, work is also being done on ‘superintelligence,’ that is, on an AI that is much more intelligent than a human.

Generative AI: AI technology capable of generating new texts, images, codes, and much more. Think of all the interesting (although sometimes problematic) responses and images you have seen produced by ChatGPT or Google’s Gemini. Generative AI tools work with AI models that are usually trained with large amounts of data.

Hallucinations: Since generative AI tools are only as good as the data they have been trained on, they can confidently ‘hallucinate’ or invent what they believe to be the best answers to questions. These hallucinations mean that systems can make factual errors or give incoherent answers. There is even some controversy over whether AI hallucinations can be ‘fixed.’

Bias: Hallucinations are not the only problems that have arisen when dealing with AI, and this could have been predicted – after all, AI is programmed by humans. As a result, depending on their training data, AI tools can demonstrate biases.

ChatGPT DOWNLOAD

AI Models: AI models are trained with data so they can perform tasks or make decisions on their own.

Large Language Models, or LLM: A type of AI model that can process and generate text in natural language. Claude, from Anthropic, is an example of an LLM, which is described as ‘a helpful, honest, and harmless assistant with a conversational tone.’