Richie Etwaru is Co-founder & CEO at Mobeus. He’s also a former CTO, CDO & CIO at Fortune 500 firms in Financial Services and Healthcare.
Although generative artificial intelligence (GenAI) has been making headlines, another of today’s most tantalizing and controversial topics is the concept of artificial general intelligence (AGI). The idea of AGI—a machine with “the ability to understand, learn, and apply knowledge across a wide range of tasks,” much like a human—has captured the imaginations of scientists, entrepreneurs and science fiction writers alike. However, despite the allure of creating such a machine, a growing body of evidence suggests that AGI will never be realized.
The Nature Of Human Intelligence
Human intelligence is fundamentally collective and constantly evolving. As individuals, we contribute to a vast pool of knowledge that grows exponentially over time. This collective intelligence isn’t merely the sum of all human knowledge but a complex, interconnected web of ideas, insights and innovations that continuously build upon one another. I’m deliberately excluding human instincts from the dialogue, as this requires another article.
This compounding nature of human intelligence presents a significant challenge for the development of AGI. To achieve parity with human intelligence, an AGI would need to encompass the full breadth and depth of this collective knowledge and evolve at the same pace. The sheer scale of this task is daunting, as it would require not only the accumulation of vast amounts of information but also the ability to understand and integrate it in ways that mirror the dynamic, ever-changing nature of human thought.
The Boundaries Of Machine Learning
Current advancements in AI, particularly in machine learning, highlight the limitations of creating a truly general intelligence. Machine learning models, such as those used in natural language processing and computer vision, excel at specific tasks by learning patterns from large datasets. These models can be incredibly powerful, achieving superhuman performance in narrowly defined domains. However, their learning process is fundamentally different from human learning.
Machine learning relies on identifying and extrapolating patterns from data humans have labeled as correct. This process can be thought of as “stretching” the confirmation provided by humans to then make educated guesses about similar data. However, this stretching isn’t equivalent to the kind of learning that occurs in humans. It’s more akin to sophisticated pattern recognition, and the further the model stretches from the original human confirmation, the greater the likelihood of error and imprecision.
The Complexity Of Human Cognition
Human cognition is a multifaceted and deeply intricate process that involves not only logical reasoning and pattern recognition but also emotional intelligence, creativity and social understanding. These aspects of human intelligence are deeply intertwined and contribute to our ability to navigate complex social environments, solve novel problems and generate innovative ideas.
Research in cognitive science and psychology underscores the complexity of human thought processes. For instance, studies on human creativity reveal that it involves a unique combination of divergent thinking (generating multiple ideas) and convergent thinking (narrowing down to the best idea)—processes that aren’t easily replicated by machines. Similarly, emotional intelligence, which encompasses the ability to recognize, understand and manage our own emotions and those of others, is a critical component of human intelligence that remains elusive for artificial systems.
The Role Of Embodiment In Intelligence
One of the key arguments against the feasibility of AGI is the importance of embodiment in the development of intelligence. Human intelligence is deeply rooted in our physical experiences and interactions with the world. This concept, known as embodied cognition, posits that our cognitive processes are shaped by our physical bodies and the environment in which we operate.
Artificial systems, lacking a physical body and the rich sensory experiences that come with it, face significant challenges in developing the kind of understanding humans possess. Without the ability to physically interact with the world, an AGI would be deprived of the context and grounding essential for true comprehension and learning.
Insights From Key Thought Leaders
Prominent thinkers and researchers have weighed in on the challenges and limitations of AGI. For instance, renowned cognitive scientist Marvin Minsky, one of the pioneers of artificial intelligence, acknowledged the difficulties of creating machines with human-like intelligence. He noted that human intelligence isn’t a “single, monolithic capability” but a collection of diverse and interdependent skills and processes.
Similarly, philosopher John Searle, known for his work on the philosophy of mind, has argued that machines, regardless of their computational power, lack the intrinsic understanding that characterizes human cognition. His famous Chinese Room argument illustrates that syntactic manipulation of symbols (which machines do) isn’t equivalent to semantic understanding (which humans possess).
The Evolutionary Perspective
An evolutionary perspective further underscores the implausibility of AGI. Human intelligence has evolved over millions of years, shaped by countless environmental pressures and genetic variations. This evolutionary process has resulted in a highly adaptive and flexible form of intelligence, finely tuned to our specific needs and circumstances.
Creating an AGI would require replicating this intricate and lengthy evolutionary process—a feat that seems improbable given the current state of technology. Moreover, human intelligence isn’t static; it continues to evolve in response to new challenges and experiences. Any AGI would need to not only match but also keep pace with this ongoing evolution—a task of staggering complexity.
The Impossibility Of AGI
Although the dream of creating an AGI continues to inspire and motivate researchers, the overwhelming evidence suggests that such a goal is unlikely to be achieved. Human intelligence is a unique and multifaceted phenomenon that arises from our collective knowledge, cognitive complexity and embodied experiences. The limitations of current AI technologies, coupled with the profound challenges of replicating the evolutionary processes that shaped human intelligence, make the prospect of AGI highly improbable.
As we continue to advance in the field of artificial intelligence, it’s crucial to recognize and appreciate the unique qualities of human cognition. Rather than striving to create machines that mimic our intelligence, we should focus on developing technologies that complement and enhance our capabilities, fostering a future where machines focus on being helpful to humans instead of being as or more intelligent than us.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
