What are the Three Types of AI? Exploring Narrow, General and Super Intelligence

For many people, AI has become shorthand for anything computer generated or assisted. The reality is a little more nuanced than that. AI has been around for some time now, even though we’re actually a long way from reaching its full potential. In this blog we’ll take you through the AI story, exploring three key phases of its evolution – narrow, general and super intelligence.

Artificial intelligence (AI), so central to the public discourse today, can actually be quite a nebulous term when you ask people what it actually means. In our previous AI blog post we defined AI, taking you back to Alan Turing in the 1950s when the question ‘can machines think?’ was first tested. Since then we’ve come a long way, to the point where ‘AI’ is often used as shorthand for anything computer generated or assisted – including content, videos, and even popular music.

AI is actually a very broad field – incorporating the likes of automation, machine learning, natural language processing (NLP), large language models (LLM), and generative pre-trained transformers (GPT) – but in tracing its journey we can divide AI into three broad types – artificial narrow intelligence, artificial general intelligence, and artificial super intelligence.

In this blog we’ll explore the AI story through the lens of these three types of AI: narrow, general, and super intelligence.

A Short History of AI

The concept of AI actually dates back to antiquity, when Greek myths spoke of self-operating machines and artificial beings. But it was the 1950s when AI really started becoming a reality.

• 1950s. Alan Turing proposed the Turing test, and the first two AI programmes – one that played chess and chequers and the Logic Theorist which performed automated reasoning.
• 1960s. The first industrial robot, Unimate, worked on the assembly line at a General Motors factory in New Jersey.
• 1970s. Political scientist Herbert A. Simon won the Nobel Prize in Economics for his theory of bounded rationality, a foundational element of AI.
• 1980s. Stanford University hosted the first ever National Conference of the American Association for Artificial Intelligence (AAAI).
• 1990s. The ‘90s saw huge advances in AI. Famously in 1997 the Deep Blue chess machine developed by IBM defeated then world chess champion Garry Kasparov.
• 2000s. More rapid progress, including Google building an autonomous car in 2009.
• 2010s. Apple’s Siri, Google’s Google Now and Microsoft’s Cortana were launched as smartphone apps that used natural language to help users.
• 2020s. ChatGPT becomes ubiquitous worldwide, leading to a surge in hype around AI and heightened discourse around the opportunities – and risks – it presents.

It’s fair to say we’ve come along way since the 1950s – from mediocre chequers playing programmes to changing society as we know it. But despite all of this progress, we’re still only in the first stage of AI: artificial narrow intelligence.

Artificial Narrow, General, and Super Intelligence

The three key phases of AI’s evolution are narrow, general, and super intelligence. Here’s a brief introduction to each:

• Artificial narrow intelligence is where we are today. Also known as weak AI, artificial narrow intelligence is designed to perform specific tasks and is not able to learn things beyond the tasks it is programmed to perform. Examples of artificial narrow intelligence include the aforementioned Siri and self-driving cars, and also the way streaming platforms like Netflix and Spotify leverage usage data to tailor recommendations to individual users.
• Artificial general intelligence is where we are heading. Also know as strong AI, artificial general intelligence is designed to perform a range of tasks in the same way as a human would. Artificial general intelligence has a huge number of potential use cases in industries like healthcare, where AI could help professionals reach accurate decisions much faster, logistics, where transportation could be made significantly more efficient, the law, where access to legal support could be made open to many more people, and almost any industry you care to think of.
• Artificial super intelligence is what many people are most excited about – and fearful of. It’s also a way off. Artificial super intelligence will outperform human intelligence, and both its proponents and its detractors believe it has the potential to change the world as we know it. We have to look to the realms of science fiction, such as R2D2 from Star Wars and Data from Star Trek, to provide examples of artificial super intelligence, which should give an indication of just how far off it is.

Understanding the Three Types of AI

In this blog we’ve explored the history of AI, along with the three key phases in its evolution. Right now, we’re in the early stages of the general intelligence phase, and it’s fair to say that super intelligence is some way off.

There are multiple uses cases for AI in organisations today. However, to get the most from AI your organisation must ensure it has the stable foundations of usable, optimised data – AI is only ever as good as the data it has to work with. Speak to one of our experts about how your organisation can take steps towards building AI into its go-forward strategy.