Humanity has consistently positioned itself at the centre of the ongoing conversation about intelligence and consciousness, frequently ignoring the complexity of other species and even more recent developments in AI. This viewpoint, which is generally known as speciesism, emphasises the innate human tendency to think of our species as superior in many ways, including intelligence. Large language models (LLMs) like ChatGPT have recently been released, reviving this discussion.
Speciesism: A Deep-Seated Belief
Speciesism is not a fleeting concept; it’s deeply ingrained in human thought. It’s not merely a software in the human mind but an immutable aspect of our mental hardware. This philosophical stance has perpetuated the idea that humans stand unparalleled in intelligence and consciousness, overshadowing the understanding of other species’ intellect.
Scientific research has gradually started challenging this anthropocentric view, recognizing that different species have evolved varied forms of intelligence based on their specific environments. Whether it be on land, in water, or the air, the arrangement of intelligence varies widely. The comparison of intelligence between different species, such as humans and cats, often proves arbitrary and unproductive. An illustrative example is that cats might have a better survival chance on a desert island compared to a human.
The Advent of LLMs: A New Dimension
The rise of LLMs introduces an entirely different type of intelligence—one that exists in a digital environment. Unlike biological beings that rely on sensory experience to model the world, these non-material forms of intelligence operate on principles that are largely incomprehensible to human understanding. This stark difference only exacerbates the speciesism issue as we attempt to compare this new form of intelligence with our own.
A recent publication by Roman Yampolsky on existential risks related to AI outlines the problem with this human-centric approach. The questions raised therein, such as whether AI will become “smarter” than humans, exemplify how our anthropomorphic lens limits our ability to understand the distinctiveness of AI intelligence. Applying a universal human ruler to measure LLMs’ likelihood to “surpass humans” only leads us into a conceptual dead-end.
Beyond Human-Centric Concepts
The inherent difference in the way LLMs function calls for a shift away from traditional human-oriented concepts like mind and consciousness. This shift must also extend to our understanding of other species, recognizing that they are not inferior to humans from an evolutionary standpoint.
Cats, dolphins, elephants, crows, and many other species possess forms of intelligence that may even be considered superior in certain aspects. Rather than engaging in a futile comparison of intelligence across species, our focus should be on appreciating the diversity and complexity of intelligence as it manifests in different life forms and even in artificial constructs like LLMs.
The speciesism conundrum shapes our understanding of intelligence in a rapidly changing world; it is more than just a philosophical argument. The emergence of LLMs has increased the need for us to reconsider our anthropocentric viewpoints and acknowledge that intelligence is complex and not limited to human characteristics.
A humble understanding of the distinctive qualities and potentials of different forms of intelligence—whether animal, human, or artificial—becomes paramount as we continue to develop and interact with increasingly complex AI models. The capacity to accept this complexity might open up fresh viewpoints and lines of inquiry, taking us beyond the boundaries of human-centric comparisons and towards a richer comprehension of intelligence in all its varied manifestations.
The article was written with the support of the telegram channels https://t.me/theworldisnoteasy and https://t.me/ai_newz.
Source: mPost
