Making sense of AI by grouping into capability and functionality types
Artificial intelligence (AI) dominated headlines in 2023. Its quick rise into the mainstream has sparked a wide range of emotions, from fear to excitement, leaving some to wonder whether technology might be humanity’s undoing.
A little concern is natural, but the idea of machines rising up, Terminator-like, and waging war against mankind is pretty farfetched. For all its remarkable abilities, AI is nowhere close to being self-aware – and despite the warnings of apocalyptic sci-fi Hollywood blockbusters, probably never will be.
Brief history of AI
AI feels futuristic, but it has been around a long time. Its origins can be traced back to ancient myths and legends. Like the automaton Talos in Greek mythology and the magical golem creature in Jewish tradition, the idea of artificial beings with intelligence has captivated imaginations for thousands of years. Interest has grown as technology has advanced.
British mathematician and philosopher Alan Turing’s seminal paper, “Computing Machinery and Intelligence,” published in 1950, proposed a test to determine if a machine could exhibit intelligent behavior equivalent to a human. This laid the groundwork for modern research into artificial intelligence.
Early breakthroughs in the 1950s and 1960s saw advancements in machine learning, natural language processing, and game playing programs. The limited computational power of the 1970s along with unrealistic expectations led to a decline in AI research.
Advances in computer power and the rise of machine learning techniques revived AI interest in the 1980s. Significant progress in areas like image recognition, natural language processing, and robotics led to a sort of second renaissance in AI. Today, it’s woven into the fabric of our daily lives, from the recommendations on your favorite streaming service to the spam filter protecting your inbox.
Because artificial intelligence is a vast and rapidly evolving field, it can be confusing to wrap your head around what exactly AI is and how it all works.
One way to make sense of this complex landscape is through categorization. Just like books, music, and film are organized by genre, categorizing AI can help us understand its different capabilities and approaches.
AI can be categorized in several ways. Two of the most common are by capability (intelligence) and functionality (learning approach). Here’s what that means:
- Narrow AI (Weak AI). This is the most common type seen today. It focuses on performing specific tasks very well, like playing chess, recognizing faces, or translating languages. Narrow AI can’t learn or adapt beyond its programmed function.
- General AI (Strong AI). This theoretical AI would possess human-level intelligence and be able to understand and learn any intellectual task that a human can. It’s still in the realm of science fiction, but some researchers believe it could be achievable in the future.
- Superintelligent AI. This hypothetical AI would surpass human intelligence in all aspects, potentially posing existential threats or ushering in a new era of technological advancement. Its existence is purely speculative at this point.
- Reactive machines. These AI systems react to the present environment without considering past experiences or future goals. Examples include chatbots, spam filters, IBM’s Deep Blue chess-playing computer, and Netflix’s recommendation engine.
- Limited memory machines. This type of AI system stores and utilizes past information to adapt its behavior accordingly and improve its response over time. Think image recognition, language translation, and self-driving cars.
- Theory of mind AI. This hypothetical AI would understand the thoughts and emotions of others, allowing it to interact on a deeper level. Though far-fetched, research in emotion recognition and social intelligence could pave the way someday.
- Self-aware AI. This even more futuristic AI would possess consciousness and self-awareness, capable of understanding its own internal state and thoughts. It’s purely conceptual at this point and raises profound ethical and philosophical questions.
Whether and how to achieve human-level or even superintelligence remains an open question, driving ongoing research and debate. It’s important to note that so little is known about how the human brain works – particularly when it comes to memory, learning, and decision-making – that being able to create an artificial version that mirrors natural biology seems highly unlikely.
For now, we’ll just leave that up to the James Camerons of the world.
Sources: Forbes.com, techtarget.com, and IBM.com