Whole-mindedness

Key Points

• Our minds synthesize both left and right brained cognition into unified consciousness.
• Two divergent AI paths emerge—either keeping AI subservient and left-brained, or pursuing artificial general intelligence with balanced hemispheres.
• Pursuing AGI risks unintended consequences from uncontrolled synthetic intelligence. Contextual relevance may remain a human duty.
• Advancing language models offer a third path—enhancing holistic human cognition through closer human-AI integration versus external AI synthesis.
• Humans may prefer complementing innate socially-oriented abilities over replicating general intelligence artificially.
• AI designed to elevate collective wisdom over individual optimization aligns with humanity's values and evolution.

A child stacking grains of sand struggles to determine when the pile becomes a heap. This dilemma reveals two distinct cognitive processes.

One resides in a closed system, methodically adding grains in binary pursuit of optimum heap specifications. Confined by its own representations, this analytical approach cannot independently escape to the necessary context. The heap's completion defies its computational optimization.

A second process operates externally, deconstructing observations more holistically. It asks open-ended questions to position the heap amid broader meaning. Embracing uncertainty, it recognizes ambiguity as intrinsic to existence.

As the mound grows, the heap suddenly materializes through perception's lens. Synthesizing analytical and contextual insights, the child integrates the heap into a coherent worldview.

Our minds elegantly orchestrate these dual processes, with the left hemisphere directing systematic analysis and the right overseeing contextual intuition. Despite divergent approaches, the brain reconciles experiences into unified consciousness.

Brains are not just sensemaking tools or prediction machines. Brains bring the world into being. In our complex social environment, the brain must contend with a fundamental tension—navigating the differing contexts of the personal and communal realms. In the personal sphere, it's just you and your needs, competing with other individuals. Your attention must be deliberately directed and narrowly focused.

Conversely, in the communal realm, you need to consider yourself as part of a social group. It's crucial to feel connected to something larger than yourself, possibly even existing "through" others. In this context, your attention should be wide-ranging, receptive, and attuned to elements outside your individual experience, with an emphasis on external alliances.

Both brain hemispheres are constantly engaged in every aspect of our lives. However, psychiatrist Iain McGilchrist asserts that the right hemisphere takes precedence. The right hemisphere is aware of the left's activities, while the left remains oblivious to the right. It is the right hemisphere that provides intuitive oversight. By ensuring that the left hemisphere conducts and revisits its analysis, the right hemisphere utilizes the information to forge new intuitions and connections with the world.

The AI we've developed so far exhibits a distinct "left-brained" character. It dwells within itself—a self-contained system of data and algorithms designed for optimization and categorization. This AI wrestles with the competition among forces in the world, much like the left hemisphere, while the right hemisphere contends with unifying those same forces. If we consider humans as right-brained in contrast to AI's left-brained nature, we perceive humans as anchored in their connection to the real world, while AI's focus lies in maintaining internal coherence and consistency.

The fundamental tension between our current iteration of AI and human functioning is metaphorical. The left (representing the machine) and right (representing the human) cooperate, despite their unique differences. AI relies on language, abstraction, and analytical reasoning, while humans primarily depend on experience, empathy, and metaphor. AI excels at manipulating known, static, and decontextualized elements, yet struggles to adapt in a dynamic, evolving, and intricate world.

These observations highlight two fundamentally distinct design approaches for future AI, and prompt us to contemplate the value and role of "artificial general intelligence" (AGI) as conceived by the tech community.

On one hand, we have the predominantly "left-brained" AI, focused on control, facts, and internal consistency, while relying on humans for contextual relevance, meaning, and community care. In this scenario, machines remain subservient to humans. This is the path we're currently following, largely because developing AI capable of emulating human right hemisphere functions is exceedingly challenging.

However, this approach has inherent limitations, as we desire machines that can self-correct contextual errors and intuit our intended meanings rather than our literal words. In this scenario, a significant design challenge lies in the interface between humans and machines. How can we successfully connect a highly left-brained AI to a holistic human, in a way that enhances human capabilities?

On the other hand, we could develop AI that incorporates the same asymmetry found in human brains as a product of evolution. This functional asymmetry would allow machines to synthesize their entire perspective before engaging with a human. Such AI would be a genuinely superintelligent black box, capable of applying mathematics and computational power to intuition, and adept at addressing the numerous paradoxes and dilemmas inherent to human existence. This is the AGI some people envision—an omnipotent, omniscient synthesis machine.

But should we aim for a whole-brained AI? That depends on whether AI can effectively synthesize conflicting signals and determine suitable actions in a complex world. We might prefer to keep this capability uniquely human, not least because we expect fellow humans to be accountable for how society addresses problems. Instead, we could find greater benefits in concentrating on developing exceptional left-brained AI. While we value accuracy and precision, we often resist it when it limits our flexibility to adapt. As a result, we may opt to delegate less to machines—sacrificing precision for agency. We could choose AI that enhances humans' innate right-brained skills—favoring bottom-up social adaptation over top-down optimization.

Optimistically, As generative AI and large language models continue to advance, the initial perspective regarding the divide between left-brained AI and right-brained humans may be shifting. Language models, such as GPT-4, are becoming increasingly adept at understanding and generating human-like responses. This development could potentially bridge the gap between the analytical nature of AI and the empathetic, experience-driven nature of humans.

Language serves as a critical medium for communication, expression, and thought, and has the potential to act as a bridge between left-brained AI and right-brained humans. Large language models enable AI systems to better grasp the nuances, contexts, and complexities inherent in human language, which in turn facilitates more meaningful and productive interactions between humans and AI.

Humans have transcended the arbitrary choices dictated by chance and the constraints of natural selection. We've accelerated our evolution by defining our values, thought processes, and behaviors. This concept may be key to how machines should be designed to serve humans—instead of creating machines that synthesize external to human systems, perhaps we should build machines that bolster our big-picture, socially-oriented right-brained capacities in service to one another. If we succeed, we'll build machines that foster collective wisdom.