Fifty years ago, we learnt that AI works better on narrow, specific problems. Problems that we would expect a child to solve, that required the application of contextual knowledge, proved difficult for AI. Although machine learning is more flexible than the older rule-based systems, it still performs much better when dealing with specific problems and is much more likely to generate incorrect answers known as hallucinations when faced with less well-specified problems.
This leads to a distinction between Narrow AI (also known as weak AI) and General AI.
Narrow AI, also known as Weak AI or Artificial Narrow Intelligence, refers to AI systems that are designed to perform a specific task or a set of closely related tasks. Unlike General AI, which aims to replicate human intelligence and perform any intellectual task that a human being can, Narrow AI is limited in scope. It operates under a pre-defined set of rules and cannot exhibit the same level of understanding or adaptability as a human.
Narrow AI is the most common form of AI that we encounter today. It is programmed to perform singular tasks such as facial recognition, language translation, or playing chess, and it does so with proficiency often surpassing human capability. However, it lacks consciousness, genuine understanding, and the ability to apply knowledge to different contexts beyond its specific programming.
The capabilities of Narrow AI are impressive within their domain. These systems can analyse large datasets more quickly and accurately than humans, identify patterns and trends, and make data-driven predictions or decisions. They are also capable of learning and improving over time through techniques such as machine learning, particularly deep learning, where they can adjust their algorithms based on the data they process.
However, the capabilities of Narrow AI are bounded by their predefined functions. They do not possess the ability to think abstractly, understand context beyond their programmed domain, or exhibit emotional intelligence. Their learning is confined to the parameters set by their algorithms and the data they are trained on.
Viewed pragmatically, narrow AI is incredibly useful. However, even since the start of the development of AI, there has been a group of people who see its purpose as replicating human intelligence. For these people, narrow AI is disappointing as it lacks many of the characteristics of human intelligence. For these people, narrow AI will always be “weak” AI and second best.
