What do you think of when you hear the words “artificial intelligence”? Maybe an image of a futuristic robot or a scene from a Black Mirror episode pops into your head, or perhaps chatbots and virtual assistants like ChatGPT and Siri come to mind. Thinking about AI might elicit some strong feelings, too; you might be excited about the advancements in AI or leery of how it could impact the future – or maybe a bit of both!
Just like us humans, AI comes in all shapes and sizes, encompassing a diverse range of forms and capabilities. To better understand and classify AI, we can identify seven general categories that can be grouped based on either their capabilities or functionalities. Let’s explore each category to understand the breadth and depth of AI’s applications and potential.
Capability-based artificial intelligence categorizes AI systems based on their intelligence. Here, AI systems are grouped according to how they learn and what they can do with their knowledge. Three types of capability-based AI exist: narrow artificial intelligence, artificial general intelligence, and artificial superintelligence.
Narrow Artificial Intelligence
Also referred to as weak AI, narrow AI (NAI) operates within a specific task and uses a learning algorithm to accomplish its objectives. Unlike broader AI systems, the knowledge acquired by NAI is not applied to other tasks. Despite their limited scope, narrow AI devices have become an integral part of our daily lives – you may have even used one to find this blog post! Narrow AI devices include Google Search, Google Home, Siri, and Alexa.
Artificial General Intelligence
Artificial general intelligence (AGI) reaches human-level intelligence, representing the next stage of AI. This is where machines will be able to think and make decisions in a similar manner to human cognitive capabilities, potentially solving problems even better than we can.
According to Maggie Harrison of futurism.com, some experts suggest that Google’s ChatGPT is showing promising signs of evolving into a form of AGI; however, it’s important to note that current AGI developments are still in their early stages, and AI systems like ChatGPT are still heavily reliant on humans at this point.
As research and advancements continue in the field of AGI, its potential societal impact remains unknown – but there’s no doubt the realization of AGI will have far-reaching implications for our world.
Artificial superintelligence (ASI) represents a hypothetical state of AI in which computers and machines surpass human intelligence across all domains. ASI doesn’t exist yet, and time will tell if we ever reach this type of AI. For now, it remains an active area of research and speculation. While ASI may be considered a distant goal, the advancements in AI technology have undeniably pushed the boundaries of what machines can achieve. We must keep having conversations about its impact should it ever be realized, including its potential risks, benefits, and ethical considerations.
Whereas capability-based AI focuses on assessing the level of intelligence of AI systems, functionality-based AI looks at the specific functions they perform or the tasks they are designed to accomplish. In this categorization, AI systems are grouped according to their intended purpose and role in solving specific problems or addressing particular needs. The four functionality-based types of AI are reactive machines, limited memory, theory of mind, and self-awareness.
As demonstrated in the famous chess matches between Garry Kasparov and a computer, reactive machines operate based on a set range of data. These machines only respond to predefined tasks programmed by a human and therefore cannot think about the past or consider future actions beyond their immediate context. In the case of the aforementioned chess matches, the reactive machine would analyze the current state of the game, assess the available moves, and choose the best move based on predefined algorithms. However, it couldn’t learn from past mistakes or anticipate future moves.
Reactive machines rely heavily on humans to carefully define and program the specific tasks and instructions the machine will follow. While reactive machines may seem limited compared to other AI systems, they’re still useful. Examples of reactive machines in our daily lives include spam filters and Netflix’s recommendation engine. Despite their limitations, reactive machines are an important step in the evolution of AI technology.
Unlike reactive machines, limited memory AI systems can make informed decisions based on past data. While they lack the long-term memory capabilities of humans, limited memory AI utilizes short-lived or temporary memories of its past experiences in its decision-making processes. Self-driving cars rely on this relatively new AI technology to navigate roads autonomously; for example, if a self-driving car sees a pedestrian crossing the street, it can leverage its limited memory to adjust its speed and trajectory to ensure safety. As technology advances, we can expect further developments in limited memory AI, bringing us closer to AI systems that exhibit more human-like learning capabilities.
Theory of Mind
Theory of mind is a more advanced type of AI still being developed. Machines in this category will likely be used in psychology, as theory of mind AI aims to understand human thoughts, emotions, and behaviors. This advanced AI could help unravel the complexities of mental health conditions and aid in developing personalized treatments. Although the theory of mind AI is a big undertaking, the potential impact of AI systems capable of meaningful communication is significant as it could help us better understand ourselves and human interaction.
Self-awareness in the context of AI often evokes vivid images of science fiction references like the Terminator. In this stage of AI development, machines possess their own consciousness and can think and learn independently, just as humans do. Today’s AI field does not have the capability to make machines that fall into this category. Researchers and scientists will continue to explore AI technology and the possibility of self-aware AI. Still, it’s doubtful that we will see any breakthroughs in this area in our lifetime.
While everyone has their own opinion of AI, its impact on our lives is undeniable. Understanding its vast forms and capabilities allows us to embrace AI’s beneficial uses while considering what our future might look like as AI technology advances. Let’s remain informed and engaged so we can navigate the ever-changing world of AI in a way that maximizes its positive impact.
Interested in learning how Vive is utilizing AI to improve its efficiencies? Contact our team today — we’d love to chat!