The Use of Different Types of AI
The transformation brought by AI in different industries could make one think that we are close to maxing out AI’s potential, which is NOT the case. Understanding the types of AI provides a clear picture of existing capabilities and an even brighter future for AI research.
The research for opportunities to make AI emulate human functions and the degree to which an AI system can replicate human capabilities is used to determine which type of AI to implement and utilize. This means that an AI can be classified depending on how it compares to human actions in terms of versatility and performance.
Using those criteria, a system that can perform more human-like functions with equivalent levels of proficiency is considered a more evolved type of AI, while one with limited functionality and performance is considered a simpler and less evolved type of AI.
Prefer to listen to the content?
Hear ELASTECH’s CEO, Armen Tatevosian, and EVP of Global Operations, Jim Zordani, explain what makes AI uniquely suited to overcome modern cyber threats and how companies can adopt AI into their tech stacks for unparalleled protection.
Approaches to Classifying AI
Based on a factor of human resemblance, there are two general approaches to classifying AI.
Classifying AI based on Functionality
One approach is structuring AI and AI-enabled machines based on their likeness to the human mind and their functionality to “think” and perhaps even “feel” like people. With this approach, there are four types of AI or AI-based systems: reactive machines, limited memory machines, theory of mind, and self-aware AI.
Reactive machines are provided with a specific process and do not have capabilities beyond that. They are fairly simple and very efficient to implement, and they are ideally suited for tasks that require quick responses.
A current example of how reactive machines are used is navigation. Google Maps and similar applications are reactive and provide routes according to currently available traffic conditions.
The second type of AI is Limited Memory, which uses past data to make decisions. The memory of such systems is short-lived. This type uses past data for a specific period of time but cannot add to a library of experiences.
Nearly all existing applications that we know of fall under this category of AI, with common applications being chatbots and virtual assistants, fraud detection, and self-driving vehicles.
Image Recognition AI is a use case of Limited Memory that leaders are increasingly interested in. A system is trained using thousands of pictures and their labels to teach it to name the objects it scans. When a new image is scanned using Limited Memory AI, it leverages the training images as references to understand the content of the images presented and, based on its “learning experience,” accurately labels new images.
These do not have to be pictures or photos in the traditional sense, but rather digital images or document scans. Leading banks, healthcare providers, oil and gas companies, and insurers have been using this technology to reduce manual and time-consuming processes, demonstrably improving operational efficiency and increasing profit margins.
Theory of Mind
While the previous two types of AI are seen more commonly, the next two types of AI are growing as either a concept or a work in progress.
The third type of AI is “Theory of mind,” which is the next level of AI at which researchers are currently engaging in innovation. This type of AI requires a thorough understanding that the people and things within an environment can influence and change feelings and behaviors. The goal is to create an AI application that understands people’s emotions, sentiments, and thoughts. However, as mentioned before, even with many improvements in this field, this type of AI is not fully developed yet.
Possible use cases for this type of AI are in Healthcare to provide highly-personalized care. A second use case would be Education, where “Theory of mind” would be used to improve individual student engagement.
The last type of AI is self-aware AI. This is the final stage of AI development, which currently exists only in a hypothetical sense. Self-aware AI is being developed with the goal of being so akin to the human brain that it has developed self-awareness. The goal of delivering this type of AI is, at a minimum, decades away from materializing and will always remain the ultimate objective of all AI research and development efforts.
This type of AI will not only be able to understand and evoke emotions with those it interacts with, but will also have emotions, needs, beliefs, and desires of its own… And this is the type of AI that doomsayers of the technology are wary of, as they fear the unintended consequences of it. Just think about Skynet in the movie “The Terminator.” Granted, this is an extreme example, but it is plausible if AI is not appropriately governed.
Classifying AI based on Functionality
The second approach to classifying AI is by capabilities, which is used far more commonly in tech, and structures AI into three different types. These are: Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI), and Artificial Superintelligence (ASI).
Artificial Narrow Intelligence
Artificial Narrow Intelligence is the type of artificial intelligence that represents most existing AI. ANI refers to AI systems that can only perform a defined task autonomously, using human-like capabilities. These machines are only capable of doing what they are programmed to do, and they have a very limited or narrow range of competencies. All reactive and limited memory AI systems use ANI.
An example of ANI is Apple’s Siri, which operates with a limited, pre-defined range of capabilities. Siri often has limitations in capability with tasks outside of its experience.
Artificial General Intelligence
Artificial General Intelligence is the “next step” in AI research and development. AGI systems have the ability to learn, perceive, understand, and function completely like humans while independently building multiple capabilities. The promise of massive savings in time and resources provides organizations with an enormous competitive edge and drives the growing interest in AGI systems. A recent example is where AGI received a $1 billion investment from Microsoft through OpenAI.
Amongst leading experts, Artificial Superintelligence is regarded as “the pinnacle of AI research,” as ASI will become, by far, the most capable form of intelligence. In addition to replicating the multi-faceted capabilities of humans, this advanced type of AI will be exceedingly better at everything people do because of its overwhelmingly greater memory, faster data processing and analysis, as well as decision-making capabilities.
Artificial Intelligence can be classified using two approaches: Functionality and Capabilities. A deciding factor for both of these classifications is human resemblance: The more and better an AI performs human functions, the more evolved it is. Within Functionality, we have four types of Artificial Intelligence: Reactive Machines, Limited Memory, Theory of Mind, and Self-Aware AI. Is is important to note that the majority of today’s AI systems are Limited Memory applications. Within Capabilities, there are three types of Artificial Intelligence: Artificial Narrow Intelligence (ANI), Artificial General Intelligence (AGI), and Artificial Superintelligence (ASI).
While we might be far from creating machines that can solve all the issues and are self-aware, organizations are increasingly invested in understanding and determining how and where AI delivers benefits, specifically for their businesses.
ELASTECH is a company recognized for delivering game-changing AI solutions and for its commitment to fostering a deeper understanding of AI. We welcome your questions about AI and its different types and use cases, and we invite you to take a deeper dive into the practical applications of AI. Our experts will help you evaluate and confirm specific opportunities where the technology will best serve you to deliver tangible business value to your organization. Click below to schedule a free AI consultation.
Schedule your Free Consultation with an AI Expert