In my last article, Thinking about thinking: Can machines accurately self assess? , I explored whether machines have the potential to self-assess as effectively as humans. I began with the assumption that humans do, in a general sense, accurately self-assess. But that got me thinking... do we?
Of course, I offer all humans the grace of imperfection. We undoubtedly do not self-assess with 100% accuracy, 100% of the time. But as we explore the abilities of machines to self-assess, it’s worthwhile to establish a baseline: how well do humans truly measure up? How accurately can we determine and communicate our own capacities at any given time?
Thankfully, I’m not the first to ask this question. Many studies consistently confirm the finding that humans are notoriously poor at evaluating their own abilities. Whether estimating the time it will take to achieve a goal (see the infamous optimism bias, showcased in this episode of Freakonomics) or assessing general behavior, our self-knowledge is imperfect. We tend to overestimate, underestimate, or simply misjudge ourselves more often than not.
But this begs the question: why? Why are we so incredibly poor at judging our own abilities?
While some research points to cognitive biases—like the Dunning-Kruger effect, where people with lower competence tend to overestimate their abilities—there may be a more fundamental issue at play. That is, how can humans accurately assess something as nebulous, fluid, and indeterminate as the “self”? How can we assess our capacity for a task when our very understanding of the self is so malleable? As Khalil Gibran wrote in The Prophet, "For self is a sea boundless and measureless."
It is not our ability to assess that is flawed, it is the underlying issue of having nothing definitive to assess in the first place.
This perspective is shared among many philosophical giants and canonical texts. In Buddhism, the doctrine of Anatta—the concept of "no-self"—outlines that there is no permanent, unchanging self. Rather, what we call the “self” is a collection of constantly shifting physical and mental components (the Five Aggregates: form, sensation, perception, mental formations, and consciousness). From this view, any attempt at self-assessment is inherently flawed because the self is not a fixed entity—it is fluid, contingent, and bound to the impermanence of experience.
Likewise, the Upanishads—a series of ancient Indian philosophical texts—depict the self as transcending individual identity. The concept of Tat Tvam Asi, or “Thou art that,” suggests that our true self is part of a larger, boundless reality. This perspective undermines the notion of self-assessment as being from an ego-bound standpoint, where the individual "I" that we are trying to assess may not even exist in a fixed or objective form.
With this existential ambiguity, the conversation opens up to an intriguing possibility: could machines step in where humans inherently cannot?
Machines, unlike humans, are fixed material objects. They are measurable, definable, and—at least in theory—assessable. In contrast to the fluctuating, subjective nature of human consciousness, machines operate within defined parameters. While they might currently lack the adaptive nuance of human reasoning (at least for now), they possess the unique advantage of being intricately definable.
Where humans are advantaged by subjective experience, machines are advantaged by objective measurability.
As our accepted reality morphs ever faster into what used to be science fiction, we may find that machines surpass us in ways we least expect. And as we face this possibility, perhaps it’s time to deepen our understanding of our own baseline human condition, accepting its beautiful imperfection and indefinability.
Linked in this article:
