You people utilize artificial intelligence to explore the mysteries of consciousness and quantum physics? This concept is addressed in a recently published guide. (Illustration by Freshpixel provided by Bigstock)
Will artificial intelligence be a boon for humanity, or will it lead to the emergence of a new class of conscious online entities with their own objectives?
This question has sparked numerous science-fiction narratives over the years, from “Colossus: The Forbin Project” in 1970 to “The Matrix” in 1999, to this year’s high-budget narrative about AI versus humans, “The Creator.”
The recent OpenAI management conflict, where CEO Sam Altman triumphed over the volunteer board members who had dismissed him a year earlier, has also been steeped in similar dilemmas.
If you were to categorize them into two factions, the board members of the AI community would be on the fast side, while Altman would lean towards a more aggressive approach. Additionally, there have been speculations that OpenAI might undergo a “breakthrough” that could propel the field forward rapidly—potentially too swiftly for humanity’s benefit.
Should we be worried if AI attains human-like capabilities and decides to assert its autonomy? This is just one of the issues George Musser, an experienced science writer, delves into in his newly released book titled “Putting Ourselves Back in the Equation.”
To delve into one of life’s profound enigmas— the nature of consciousness—Musser engaged with experts in AI, quantum physics, neuroscience, and philosophy. Is consciousness an exclusively human trait?
What was his conclusion? Musser suggests that the right type of AI could possess comparable awareness to humans. In a recent episode of the Fiction Science radio, Musser stated, “Almost everyone contemplating this, across various disciplines, agrees that replicating a nerve in silicon would require immense precision to mimic biology—yes, it could exhibit consciousness.”
But should we be apprehensive about the rise of dominant AI entities? Musser’s viewpoint on this existential question differs from typical science fiction narratives.
“I believe that unintelligent and unconscious models are extremely perilous—already quite capable of disrupting society—as evidenced in today’s world. He argues that conscious machines do not possess any inherent capabilities that would heighten, at least in my view, their threat to humanity.
Musser contends, “If anything, I fear we might offend them.” However, such mistreatment has been a part of human history. Until we evolve beyond that, we often mistreat beings that are evidently conscious.
While scientists researching in this domain are not particularly surprised by the capabilities of relational AI tools like OpenAI’s ChatGPT, Musser notes that they are taken aback by the rapid pace of their advancement.
He draws a parallel between the evolution of nuclear fusion and that of AI. Fusion has always seemed like a technology of the future. However, in the case of AI, the future appears to be nearer.
According to Musser, “it accelerates the timeline for what many envision as AGI.” AGI stands for artificial general intelligence, a system capable of performing diverse tasks and learning dynamically—essentially a closer emulation of an animal, human, or even a superior intellect. Thus, I believe it has hastened progress, albeit qualitatively rather than quantitatively for humanity.
George Musser, a science writer, (on the right) embraces an android named Sophia. (Photo: Courtesy of George Musser)
Musser emphasizes the growing importance for experts to comprehend how our neurons function in the era of ChatGPT. He asserts that this is not merely a scientific pursuit driven by curiosity. To evaluate machine systems and ask, “Are they conscious?,” a solid grasp of consciousness is imperative.
In his book, Musser delves into two concepts:
- Integrated information theory: This theory posits that consciousness can be quantified on a numerical scale, arising from the internal connectivity of a network. Animals possess a level of consciousness inferior to humans. While technically inanimate objects like computers could potentially exhibit awareness, their information networks would need to mirror natural nervous systems with feedback loops.
- Predictive coding: Consciousness evolved to aid organisms in forecasting future events and averting threats, such as predicting a predator’s attack. Machines might require predictive models of their environment as intricate as those developed by animals over millennia.
These are not the sole theories, according to Musser. He is currently exploring the concept of global workspaces, another intriguing idea. This theory implies that only a select few species with specific cognitive abilities are capable of consciousness.
Most scientists concur that the current generation of relational AI does not possess consciousness by any measure. Musser states, “ChatGPT, perhaps never conscious.” But what about a successor? ChatGPT 5? Possibly.
The intriguing connection between quantum physics and consciousness theories is a focal point of Musser’s work. On one hand, Nobel laureate Roger Penrose argues that the actions of neurons’ molecules underpin consciousness. If this holds true, conventional computers may not achieve genuine consciousness.
On the other hand, some researchers speculate that a deeper comprehension of consciousness could illuminate the enigmas of quantum physics, particularly the “observer effect,” which explains phenomena like superposition and entanglement. AI and the study of consciousness are gaining traction as these mysteries remain unsolved.
Regarding the peculiar role the observer seems to play in our understanding of the physical universe, Musser suggests, “specific theories in biology and philosophy of mind can aid in addressing the questions that physics poses and grapples with.”
Musser anticipates a breakthrough, not one that leads to dystopia akin to the Matrix, but rather one that yields new scientific insights comparable to Albert Einstein’s theory of relativity. Such groundbreaking discoveries may occur once in a decade over time.
He expresses excitement about the prospect of leveraging AI for this purpose. An example from radiology demonstrates how AI detects 97% of targeted anomalies like tumors or aneurysms, surpassing human radiologists’ 99% detection rate. By combining AI with human intellect to unravel the mysteries of consciousness, Musser envisions a similar enhancement.
