by Andhie Wong
Artificial intelligence is changing our world quickly, and it's worth understanding the different types of AI that get lumped together under one label. The three main categories are Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI) and Artificial Super Intelligence (ASI) — each with its own capabilities and its own impact on our lives.
Artificial Narrow Intelligence (ANI)
ANI, also known as "weak AI," is built to do specific tasks. It operates within a limited domain and can't learn beyond what it's been programmed or trained to do. We already see ANI everywhere in daily life: voice assistants like Siri and Amazon Alexa use speech recognition to set reminders or play music; recommendation systems on Netflix and Spotify analyse user behaviour to suggest films or songs tailored to individual taste; and self-driving systems from companies like Tesla use vision recognition to navigate and respond to the road.
ANI helps us across many fields. In healthcare, it can analyse patient records and assist doctors with diagnosis. In finance, it can spot fraudulent transactions.
Artificial General Intelligence (AGI)
AGI, or "strong AI," aims to think and learn across many different tasks the way a human does. Unlike ANI, which is confined to specific jobs, AGI would be able to adapt and apply knowledge across varied situations. Right now, AGI remains mostly a research goal, though many researchers are actively working to get there.
AGI hasn't been built yet, but it's a major target for AI research. Some of the ideas researchers are working toward include: truly self-driving cars that can handle any situation and adjust to unexpected challenges, requiring genuine on-the-road decision-making rather than pattern-matching; and advanced chatbots that could hold real conversations, understanding emotion and context rather than just following a script.
The search for AGI is itself a reminder of just how complex human cognition really is, and how hard it is to replicate in a machine.
Artificial Super Intelligence (ASI)
ASI is a future concept in which AI would be smarter than humans across every dimension. This kind of intelligence could, in theory, help solve major problems like climate change or disease. But ASI remains purely theoretical — no system today comes close.
If ASI were ever developed, it could unlock genuine breakthroughs in medicine and environmental science. But it also raises serious concerns: control, since people worry ASI might act in ways we can't predict or manage, making choices that don't align with human values; and existential risk, since a system smarter than humans in every way could create situations where human survival itself is at stake.
These concerns are exactly why AI needs to be developed responsibly, to make sure it benefits everyone rather than creating risks we can't manage.
Closing
As we explore AI further, it's clear we're only beginning to understand its potential. From ANI making everyday life easier, to AGI's exciting possibilities, to ASI's open questions, the journey of artificial intelligence is really just getting started. Understanding these categories helps us prepare for what's coming, and make sure AI ends up working in our favour.
Part of the AK AI Corner column. Originally published in Oriental Daily (东方日报) on Aug 17, 2024.