Artificial Intelligence (AI) is a branch of computer science that deals with the development of machines with the ability to perform tasks that are usually associated with human intelligence, such as learning and problem-solving, or put more simply, AI researchers are trying to develop machines that can think like humans. Relatively speaking, AI research is still in its infancy. In 1950, mathematician and computer scientist Alan Turing proposed a test (obviously called the Turing Test) for a machine's ability to exhibit intelligent behaviour, requiring that a human being should be unable to distinguish the machine from another human being by using the replies to questions put to both. That test has yet to be passed in any meaningful manner.
AI research often intersects with other fields of research, so it's not just one area; yes, researchers want to see if they can development intelligent machines, but then what would you do with them? Hence there's a whole network of different, but related, areas of research under the general heading of AI.
A chatbot is an app or a web platform that mimics human conversation. Chatbots are often used by websites to provide customer service. If you go on to a webpage and see a chat pop-up offering assistance, chances are it's a chatbot configured to recognise keywords in queries and give a response. The odd chatbot might pass the Turing Test, but no-one thinks chatbots are intelligent. Turing himself thought that the test simply illustrated that one day, machines might be able to convince us that they think regardless of whether they could or not. The latest generation of chatbots are just about at that stage.
In late 2022, the company OpenAI released a chatbot named ChatGPT (Chat Generative Pre-Trained Transformer). ChatGPT is much more sophisticated than other chatbots. It can (list shortened from Wikipedia):
It doesn't necessarily do these all things well, but it can do them and. what's important to note is that it can do them very quickly.
Since its release, ChatGPT has become very popular because it can respond to prompts to generate new and original text. Other tech companies, such as Microsoft and Google have accelerated their AI research, and many companies have released their own ChatGPT equivalents. These complex programs use algorithms - a defined process or set of rules to be followed in calculations or other problem-solving operations - to generate content based on conversational prompts given by humans and thus ChatGPT and similar programs are called generative AI.
Hence, generative artificial intelligence is the term used to describe computer programs, such as ChatGPT, that can be prompted to create new content, including audio, computer code, images, text, simulations, and videos. In the AI family tree above, generative AI mostly falls under Natural Language Processing (but can also be classified under speech and vision and deep learning - AI is complicated!)
The sophistication (yes, Chat GPT has passed the Turing test) of these computer programs and their capacity to process and generate human-like text has implications for content creation, conversational agents, and automation of various language-related tasks.
At the time of writing there are two versions of ChatGPT available:
These, in time, will be replaced by faster and more up-to-date models
The algorithms that generative AI programs use are quite complex and can adjust, reorganise and update according to the input they receive. This is called machine learning. As these programs use language - text - as input, they're called language models. As the datasets used are very, very large (ChatGPT was trained on 570GB of data - 300 billion words), this is expanded to large language models (LLMs). Because there's a "learning" element (revising and updating) they are sometimes referred to as Language Learning Models - the acronym LLM is the same
Generative AI is not AI as we understand it from films and television programmes like 2001, Terminator or Westworld, and when you get down to it, these generative AI programs are really quite dull. These are simply computer programs that have been "fed" (the technical term is "trained on") language data like webpages, newspapers, magazines, books and other text content. They use this data to build up a big vocabulary of words and use an algorithm to figure out words that usually appear next to other words in a given context.
The GPT in Chat GPT has been humorously described as standing for "Good Predictive Texting". Predictive texting, or autocomplete, is a feature of some apps that predicts the next text input based on the text you have entered already. For instance, in your phone's messaging app, if you start typing "grand" you might be prompted to use "grandad" or "grandma", or if you type "what is the..." you might be prompted with "time", "name" or plan" as these often follow "what is the..."
So, when you prompt generative AI, all it's doing is processing your words and estimating the probability of the words in its training data being used in the context of your conversation/prompt. It chooses one of them as being likely to be the first word in the reply and the word after that and the word after that until it has a sentence or paragraph or essay. Generative AI doesn't know anything other than what word usually comes after another and therein lies a problem; its abilities and the quality of its output are often overestimated.
Despite some limitations, generative AI has huge potential in many areas and is already has a rapidly expanding number of uses:
Image source: Amanda Wheatley and Sandy Hervieux - The AI Family Tree used under a CC-BY-NC-SA-4.0 licence
