KENNESAW, Ga. | Feb 27, 2023

The term superintelligence is provisionally defined as a kind of intelligence that surpasses human cognitive abilities in almost every area of interest. This means that any entity with intelligence greatly superior to that of a highly intelligent individual of today would be considered super-intelligent. Renowned AI philosopher, Nick Bostrom, examines various potential paths to the development of superintelligence, including advances in existing AI technologies, the creation of brain-computer interfaces and neuro-implants, brain emulation, and more [1]. This blog post briefly outlines possible paths toward superintelligence discussed in Bostrom’s book and adds two additional possibilities that are not directly covered in the book, namely the expansion of human senses and the Conscious Empathic AI paradigm. It then moves to discuss some examples of business research areas related to superintelligence and AI advancements.

Drawing upon the evolution of intelligence and the history of human technological progress, Bostrom asserts that the emergence of superintelligence is inevitable. He notes that the existence of highly intelligent humans is a recent phenomenon compared to millions of years of natural evolution. Bostrom further highlights the substantial increase in human intelligence compared to our closest animal relatives, suggesting that humans represent an explosion of intelligence on earth. Additionally, he argues that human technological advancements follow an exponential rather than linear pattern, with each new invention leading to a shorter time between breakthroughs. Based on these observations, Bostrom concludes that the emergence of superintelligence is not only possible but very likely. While we cannot be certain how it would emerge, we can have some educated speculations as outlined hereafter.

Advances in Artificial Intelligence and Machine Learning

Learning is essential to achieve general intelligence in humans, and the machine learning paradigm is a means of attaining this objective for machines. Blind evolution, a supposedly directionless evolutionary process, has led to human-level intelligence. Therefore, an artificial intelligence designed and guided by humans should, in theory, be able to achieve human-level (general) intelligence and then surpass it by comprehending its own mechanisms to engineer new algorithms and infrastructure that will enhance its cognitive performance.

Whole Brain Emulation

This approach involves closely mimicking the structure and function of the human brain to create an intelligent system. This is done by scanning and modeling the computational architecture and processes of the biological brain in great detail. The primary obstacle is how to scan a living human brain at the level of detail needed for whole-brain emulation. However, several technological advancements are necessary to achieve complete whole-brain emulation and enhance the emulated brain’s functioning and capabilities to the level of superintelligence.

Biological Cognition

This approach aims to improve the functioning of the human brain to achieve intelligence greater than that of current humans. It is a modern version of the concept of enhancing lineage and capabilities through selective breeding and nurturing. With the aid of current technology, we can optimize infant feeding schedules and contents to enhance brain development and functioning in adulthood. By combining this with tailored education, we can anticipate more intelligent humans in the future. As these more intelligent humans interbreed, subsequent generations should become even smarter, eventually leading to the emergence of super-intelligent beings leading to superintelligence.

Brain-Computer Interfaces

The integration of functionalities currently available to digital computers, such as perfect recall, fast and accurate calculations, and high-speed yet minimal error data transmissions, into the human brain could be achieved through brain-computer interfaces, mostly through implants. This path also holds promise for the development of superhumans and SI in the future. Despite many advances made by corporations focusing on this area (e.g., Elon Musk’s NeuroLink Corporation), there are still numerous health and safety issues to be addressed before these technologies and interfaces can thrive.

Networks and Organizations

Nick Bostrom’s final suggested path involves connecting individual human minds to create a larger intelligent organism to utilize the power of collective computing. Humans have historically made significant technological progress through language and connectivity, which enabled them to communicate and collaborate more effectively than other species. The goal is to further enhance this connectivity using technology and achieve “collective superintelligence.”

Expansion of Human Umwelt

The essence of this path is based on human-computer interfaces, but it differs from the path suggested by Bostrom in that it does not aim to integrate digital technology capabilities with the human brain. The idea is to expand current human senses and create new ones by introducing technologies that can sense data beyond what organic human sensors (e.g., eyes and ears) can do. The idea is that our brain can learn to create a representation of any data that is fed to it regularly and consistently. The idea has already been tested and proven to work through sensory channel conversion and expansions, where sound data is converted into a pattern of vibrations that are sensed through a deaf person’s skin [2]. This path assumes that humans already possess supercomputing capabilities, but such capabilities are underutilized; If we can feed more data of novel types and expanded variations to our innate supercomputers (brains), in time, we can become super intelligent.

The Brain Lab and the Immersive Visualization Environments Research Cluster at the Coles College are the two research centers containing the infrastructure that, with the help of some additional technology, can enable Coles scholars to conduct research related to human umwelt expansion and its effect on individuals and organizations. One interesting area is to consider feeding organizational performance data directly to decision-makers instead of designing dashboards and written reports.

Conscious Empathic AI

The final approach can be viewed as a variation of the first but differs in its ultimate objective. Creating a machine that can achieve linguistic indistinguishability from humans and pass the Turing Test has been the primary objective of AI and machine learning research and practice, and passing this test is seen as the first major step toward the creation of machines with general intelligence and phenomenal consciousness [2]. With the introduction of ChatGPT by OpenAI and the upcoming rival services, one can argue that the Turing Test is already passed and the long-standing goal of the AI field has been achieved. There is an intelligent entity that can write quality poems, compose music, answer almost any kind of question, and write computer codes, essays, and articles in the English language. It does all that in a way that most English-speaking people cannot easily distinguish whether a human or a machine has produced the writing. Considering the basic definition of superintelligence: “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest” [1. p26] one may even argue that we already have SI since ChatGPT performance exceeds that of average humans in almost all cognitive and intellectual areas!

On the other hand, despite mesmerizing advances in AI research and practice, the conscious empathic AI paradigm contends that for AI to advance to the next level of developing metathinking, creativity, and empathy, it first needs to become conscious following a similar pattern to that of human evolution. To that end, it proposes that “consciousness in AI is an emergent phenomenon that primordially appears when two machines co-create their own language through which they can recall and communicate their internal state of time-varying symbol manipulation.” [3 p.549] In this view, consciousness arises from the communication of machines’ inner states. The ability of individuals to communicate their inner state to others and the other’s capability to internally reconstruct such communicated states is at the center of human empathy. Hence, it is expected that the machines’ ability to communicate and reconstruct inner states also leads to the rise of empathic qualities. Consequently, machines capable of communicating their internal states using their own complex language and collaborating with each other to accomplish various goals - not included in their initial programming - would also be able to understand their internal workings and improve them to achieve superintelligence.

Future of Business Research

Some reputable business scholars believe that after taking over lower-level mechanical jobs, AI is now well-positioned to take over many analytical jobs that involve critical reasoning and decision-making—essentially making it suitable to take over middle tactical management and supervisory jobs [4, 5]. As business researchers, we should start considering the effects of such job shifts on individuals and organizations. There are many interesting areas of study at both micro and macro levels. For example, business scholars might consider questions such as: How would replacing human managers with AI impact perceived procedural and distributive justice in organizations? How would robo-managers impact employees’ job satisfaction or organizational performance? Or how does the extensive use of AI in managerial positions change the organizational structure? Will it make organizations flatter or more hierarchical? And how would such a change impact organizational performance?

When it comes to the prospect of Conscious Empathic AI emergence, we face even more fundamental questions. For example, a pivotal consideration is whether conscious AI agents can be unplugged, deleted, or decommissioned. Does the carbon-based consciousness (humans) have an inherent right or superiority over the potential silicon-based consciousness? The question of the conscious AI’s right to exist is fundamental and must be philosophically explored before the advent of AI consciousness. Scholars should also examine what it means for conscious AI agents to serve humans. And in that vein, various issues concerning conscious AI agents’ rights and the relevant service laws, regulations, and policies need to be discussed [3]. For instance, should conscious empathic AI agents be entitled to a share of the value they co-create by serving humans, and should labor laws be adjusted to protect their rights? Theoretical discussions must start, which may eventually lead to legislation clarifying responsibilities and accountabilities. New research would be needed to understand the agentic relationship dynamics between humans and conscious empathic AI to address situations where service machines may act against the interest of their owners, operators, or clients. Aside from these fundamental issues, the emergence of conscious AI would impact most of the AI in business research centered on humans’ anthropomorphic perception of machines, requiring anthropomorphic-centered research assumptions to be revisited and reexamined.

Finally, the prospect of the superintelligence phenomenon is no longer purely reserved for sci-fi books and movies. Such a prospect is backed by deductive logic based on the history of human evolution and technological advancements. Even though how and when it would happen is subject to debate, Its occurrence is almost certain. Hence, business scholars should start discussing and theorizing how superintelligence will change individuals and organizations, depending on the path or combination of paths through which it will be realized. Superintelligence realized through the “networks and organizations” path will have a drastically different impact on humans than one realized through advancements in machine learning, brain implants, or a combination of suggested paths.

We are living in exciting times. We are creating something fundamentally different from all of our past innovations and creations. Essentially, we are succeeding in creating something in our image! And as a result, an utterly uncharted territory is slowly opening up and waiting to be explored and mapped through our theorizations and experimentations!

References:

1 - Bostrom, N. (2015), Superintelligence: Paths, Dangers, Strategies. New York, NY: Oxford University Press.

2 - Bringsjord, S. and Govindarajulu, N. S. (2018), “Artificial Intelligence,” in The Stanford Encyclopedia of Philosophy, E. N. Zalta, ed. Stanford, CA: Metaphysics Research Lab, Stanford University.

3 - Esmaeilzadeh, H., & Vaezi, R. (2022). Conscious Empathic AI in Service, Journal of Service Research, 25(4), 549–564.

4 - Huang, M. H., Rust, R., & Maksimovic, V. (2019). The feeling economy: Managing in the next generation of artificial intelligence (AI). California Management Review, 61(4), 43-65.

5 - Rust, R. T., & Huang, M. H. (2021). The feeling economy: How artificial intelligence is creating the era of empathy. Cham, Switzerland: Palgrave Macmillan.

— The decorative picture is generated using the OpenAI text-to-image generator

DALL-E.