Understanding Artificial Intelligence: Its Definition and Boundaries

Artificial Intelligence (AI) is a widely used but often misunderstood concept. While exploring this field, I've found it essential to understand what AI truly involves and what it does not.

What is AI?

Artificial intelligence (AI) is commonly understood as a technology that simulates human intelligence. This includes exhibiting abilities such as learning, problem-solving, decision-making, reasoning, perception, comprehension, and understanding of language, images, sounds, and more.

AI is also recognized as an enabler, creating and powering systems, machines, devices, and appliances to perform tasks that typically require human intelligence.

Through my investigations, I have come to understand that for a technology to be considered AI, AI-driven, or AI-powered, it must demonstrate specific characteristics:

• An ability to execute an input algorithm.
• An ability to solve problems; process, predict and decide based on the data.
• An ability to process data; collect, filter, analyze, and/or transmit data.
• An ability to predict based on data processed; anticipate action and its effectiveness.
• An ability to decide; take action.
• An ability to learn and be trained.
• An ability to adapt; and apply learnings to other situations.
• And an ability to do all the above autonomously, without human intervention.

AI is a technology that enables systems to use algorithms to solve complex problems, make accurate predictions, make informed decisions, and improve itself through learning and adaptation. All done autonomously.

Some well-known examples of AI-powered systems include chatbots, self-driving cars, and robots.

To understand AI, it's important to distinguish between the two main types of AI systems – narrow AI and general AI.

What are the main types of AI?

Narrow or Weak AI refers to specialized artificial intelligence designed to perform specific tasks or solve particular problems, such as speech recognition, recommendation algorithms, and autonomous driving. These AI systems excel in their designated functions but cannot perform tasks beyond their programmed capabilities. Some examples include Siri, Alexa, recommendation algorithms on streaming services, image recognition software, and most virtual assistants.

General AI or Artificial General Intelligence (AGI) is where it can get exciting. These are AI systems that can mimic human intelligence across various tasks, with the ability to understand, learn, and apply knowledge across a wide range of functions at a level comparable to human intelligence. AGI can hypothetically perform any intellectual task that a human can, adapting and applying knowledge from one domain to another. The concept of AGI is still in its early days and presents significant scientific and ethical challenges.

What are the component capabilities of AI?

The concept of AI encompasses a set of component traits and capabilities that are expected to be demonstrated by intelligent systems. Some of the primary capabilities include the following:

• Machine learning involves training algorithms on large datasets to recognize patterns and make predictions and decisions.
• Deep learning is a more advanced subset that uses neural networks with many layers to analyze vast amounts of data, achieving remarkable accuracy in complex tasks like image and speech recognition.
• Natural Language Processing (NLP) understands and processes human language (e.g., chatbots, translators).
• Computer Vision enables machines to interpret visual information (e.g., facial recognition, self-driving cars).
• Robotics combines AI with mechanical systems to perform physical tasks.
• Expert Systems mimic human expertise in specific fields (e.g., medical diagnosis).
• Reinforcement Learning learns by trial and error to achieve specific goals (e.g., game playing).

What is not AI?

AI is often confused with general-purpose software automation and data processing, but they are not the same. Traditional software follows explicit instructions given by programmers without learning or adapting from data. For example, a simple if-then rule-based system for sorting emails into folders is not AI, as it performs a predefined task without the ability to learn from new data or improve over time.

In addition, while AI systems can process and analyze data at incredible speeds, they do not possess consciousness, self-awareness, or emotions. Portrayals in science fiction, where AI systems exhibit human-like consciousness or intent, remain in fantasy and form part of the research into AGI. Despite their advanced capabilities, current AI technologies operate purely based on mathematical models and data-driven algorithms.

AI is also not synonymous with highly complex or sophisticated technology. For instance, traditional databases, spreadsheets, and statistical analysis tools, even when they handle vast amounts of data and perform complex computations, do not qualify as AI because they do not involve autonomous learning or adaptation.

Another misconception is equating AI with robotic automation. While AI can be integrated into robotics to enable smarter, more adaptive machines, not all robotics involve AI. Many robots operate based on pre-programmed instructions without any learning or decision-making capabilities.

In summary, AI should be distinguished from its component traits in isolation.  AI is not traditional automated software, rule-based systems or fictional depictions of sentient machines. 

AI involves systems that can utilize algorithms, process data, predict, decide, learn, and adapt with end-to-end autonomy without being explicitly programmed for every possible scenario. It's important to differentiate genuine AI from traditional automation, fictional portrayals, or transformational concepts to understand its capabilities and limitations.

In the next article, I'll delve into how AI revolutionises industries by improving operational efficiency, enhancing customer experiences, and driving innovation.        

This article is part of the series "Exploring Organizational Agility & Generative AI " authored by Chuks Anochie.

Chuks Anochie is an expert in digital transformation and technology management. He specializes in helping organizations attain organizational agility and gain a competitive edge. He has built experience over two decades in assisting enterprises aiming to enhance operational efficiencies and expedite the delivery of value.

During his downtime, Chuks engages in doctoral research within this domain and aspires to contribute to the growing academic discipline of digital transformation.