The term artificial intelligence often perplexes many, but have you ever wondered why? The root cause lies in our own struggle as "intelligent beings" to define the very essence of "human intelligence" itself. Let's unleash this enigma to enhance the right understanding of AI and its development.
The Enigma of Intelligence.
While most individuals have a general understanding of what it means to be "intelligent," precisely describing it proves to be a challenging task. Due to the concept of intelligence which is both concise and elusive.
Therefore, if you were to ask 20 intelligence researchers for a concrete definition, you would likely receive 20 different answers. Nevertheless, there is at least a consensus that intelligence manifests in various forms.
Many scientists categorize "cognitive performance" into distinct types of intelligence, such as linguistic, logical-mathematical, spatial, and social-emotional intelligence. This approach also applies to how artificial intelligence is structured, e.g.:
• NLU (Natural Language Understanding) = linguistic intelligence
• Robotics = spatial intelligence
It is important to note that these different facets of intelligence do not need to be equally developed in an individual to be considered intelligent. For instance, a so-called genius may excel in mathematics but lack social-emotional intelligence. Interestingly, however, we often expect robots to possess all forms of intelligence simultaneously and instantaneously.
In essence, both human intelligence and its artificial counterpart revolve around the ability to solve problems and process information. Humans exhibit this intelligence through their cognitive abilities, while machines demonstrate it through artificial means.
Brain vs. Intelligence.
The prevailing belief for a considerable period of time was that the brain serves as the center of intelligence. However, this assumption remains far from being comprehensively explored. The advancement of artificial intelligence has propelled brain research to unprecedented heights, with an overwhelming amount of studies being conducted in this field. Such extensive research has surpassed the capacity of a single individual to absorb all the findings. Consequently, artificial intelligence not only fuels research endeavors but also aids in the efficient dissemination of research results, enabling faster processing of data. Moreover, recent revelations have unveiled that intelligence is not solely confined to the brain, which adds an intriguing dimension to the realm of artificial intelligence. Remarkably, even seemingly "brainless" single-celled organisms exhibit intelligent behavior, offering a wealth of fascinating insights for AI development. To delve deeper into this captivating subject, I highly recommend watching the ARTE documentaries such as "The clever gut, our second brain" and "The Blop - Intelligence without a brain".
Artificial vs. Human Intelligence.
The concept of comparing artificial intelligence to humans has a rich historical background. Even in ancient Greek mythology, there were stories of "human machines" like Talos, a colossal warrior made of bronze brought to life through a "blood channel" running from head to heel.
However, it wasn't until the 1950s that the term "artificial intelligence" was coined by John McCarthy. He defined it as the science and technology used to create intelligent machines.
In the past, it seemed unimaginable to have beings more intelligent than humans. But today, this comparison falls short as artificial intelligence can outperform us in certain tasks, with greater speed and accuracy. It can also process vast amounts of data that would overwhelm us. Therefore, it is more fitting to refer to these machines as intelligent machines, as originally intended, rather than getting too caught up in humanizing them.
Intelligent machines now have the capability to take over activities that were previously exclusive to humans. However, this does not mean they have to do it in the same way, nor should their abilities be limited by comparison to humans. Comparing them to humans serves the purpose of setting goals in machine learning and identifying any inherent biases. Ultimately, the goal is not to replace humans with AI, but rather to intelligently support them. In essence, it's all about adding value.
Machine Learning vs. Artificial Intelligence.
Statements like "This is machine learning, not artificial intelligence!" or job titles like "Head of Data, AI & ML" understandably cause confusion. While artificial intelligence and machine learning are not exactly the same thing, it is not logical to strictly differentiate between them. Additionally, learning and information processing are crucial aspects of human intelligence and should not be seen as separate entities.
Machine learning is inherently a subset of artificial intelligence, encompassing areas like NLP (natural language processing) and more. Therefore, anything related to machine learning automatically falls under the umbrella of artificial intelligence.
There is often debate among professionals about whether certain machine learning algorithms are too simplistic to be considered intelligent. I disagree with this viewpoint, as even so-called "stupid" ML algorithms can be incredibly valuable. This can be compared to human intelligence, where vital functions are controlled by our "stupid" reptilian brain, such as breathing and regulating heartbeat. We would not want to disregard these functions simply because they come from a less sophisticated part of our brain. In fact, complexity does not always equate to intelligence; sometimes simplicity can be just as or even more effective.
However, it is important to note that not everything classified as artificial intelligence is necessarily machine learning. Artificial Intelligence is a broad term that encompasses various subcategories. Nevertheless, machine learning remains a significant factor within the AI field. Without machine learning, progress in other artificial intelligence areas would be mainly hindered. Just like how we don't stop to learn beyond our early childhood.
Now or Never.
The
ever-evolving realm of artificial intelligence is vast and diverse, constantly
expanding and adapting to new advancements. This dynamic nature gives rise to
an array of categories and subcategories, as innovation continues to push
boundaries. A prime example of this is the official recognition of deep
learning as a subcategory of machine learning in the year 2000, despite the ML roots dating back to 1959. In essence, the AI landscape is becoming
increasingly intricate, necessitating a firm foundation of knowledge in order
to stay abreast of long-term developments. Embracing this complexity is now
imperative, as it is the key to remaining at the forefront of this
transformative field.
Author: Carole Gächter @ ImpactOvation
Feel free to share this article and follow for more.