Intelligence Without Consciousness

Intelligence Without Consciousness

The restriction on the use of Anthropic’s new AI model, Claude Fable 5, has taken the AI debate to a new level. This is no longer a democratic tool available to all. This is a tool that may actually create a new power divide. The countries with access to, and control over, these new AI tools may eventually determine the course of the 21st century.

With AI reaching what is often described as singularity, where collective human intelligence may be surpassed by AI, the existential threat has once again become a subject of discussion. What is interesting is that these existential threats may not arrive in the way science fiction has conditioned us to imagine them—giant machines invading human civilization and destroying it. Rather, the threat may be much more subtle, as Yuval Harari explains.

I found one of his arguments particularly fascinating. He argues that the major challenge AI will bring to humanity will emerge through bureaucracy. The transformation will be subtle, yet profound. AI has the capacity to coordinate at an unprecedented scale, and it has access to language. These are the two very capabilities that ultimately established the dominance of humans over other species.

His example is striking. One tiger is more powerful than one human, but one hundred tigers are less powerful than one hundred humans. Human beings possess an unparalleled capacity to cooperate with strangers through shared stories, institutions, and bureaucracy. This capacity to organise collectively has been one of the most significant forces behind the creation and sustenance of human civilizations.

Harari argues that AI may eventually build systems of communication that are no longer comprehensible to humans. Consider what we do with chickens and cows. They share the same physical space with us, but they do not understand our coded language, our institutions, or the systems that govern their lives. The real threat, according to him, is that AI may one day construct financial systems, legal systems, or bureaucratic structures that remain beyond human comprehension.

These arguments sound both fascinating and threatening. Yet I find myself asking a different question: Why would AI do that? Is AI a conscious entity? Doesn't power and control make sense only for conscious beings? What meaning do power and control have for an entity that lacks consciousness?

A weightlifter wins an Olympic medal and enjoys the recognition society gives him/her. Machines can lift far greater weights than an Olympic champion, but does a machine gain any identity or status because of that? A powerful car may add to the identity of its owner, but the car itself possesses no identity. The value of power seems to exist only because conscious beings assign meaning to it.

In this context, I was recently reading a piece titled The Anguish of Choice:

"A person becomes who they are through others and, according to Sartre, 'he cannot be anything (in the sense in which we say someone is spiritual, or cruel, or jealous) unless others acknowledge him as such.' We discover ourselves through other people's perceptions and demands, so much so that 'we each attain ourselves in the presence of the other.' This intersubjective relationship is, in Sartre's view, the foundation of what connects humans to one another."

The real question in the AI paradox, then, is this: where is the intersubjective reality in the case of AI?

Perhaps this is the question that deserves greater attention than the fear of superintelligence itself. We often assume that intelligence naturally seeks power. But throughout human history, power has never been an end in itself; it has been meaningful because conscious beings desired recognition, status, identity, and belonging. If AI does not possess consciousness, desire, or an intersubjective existence, then our assumptions about its motivations may require as much scrutiny as our assumptions about its capabilities. The future of AI may therefore depend not only on how intelligent machines become, but also on whether intelligence without consciousness can ever become an independent agent of history.