Will AIs Follow Corporations Into Personhood?

Speculation about whether or not AI is conscious is reaching a fevered pitch on social media and the blogosphere. The problem is that everyone is talking past each other because there is no agreed-upon definition of consciousness and people are not stating from what ontological paradigm they are arguing.

Getting an answer to the question “is AI conscious” seems to become more urgent because many of the abilities that were once seen as indisputable proof of consciousness are now common in frontier AI models. Concerns are arising about welfare and personhood of AIs.

In an earlier post I attempted to define a number of axes of consciousness based on Buddhist teachings so as to clarify the difference between human and AI consciousness. I laid out the axes on a radar plot like so:

This chart illustrates the idea that consciousness is not some kind of homogeneous substance but that it is, perhaps somewhat like intelligence, a varied composite of multiple components.

Next, let’s look at how the prospect of AI consciousness is evaluated based on some common and not so common philosophical frameworks. I am including the less common filter model that I subscribe to myself.

Physicalism

This is the most commonly accepted view in science. The idea is that everything is physical and that there is nothing above or beyond the physical. Based on this view consciousness arises from complex information processing. Within physicalism some believe a biological substrate is required for consciousness while others believe it could arise independent of substrate. For the latter group AI consciousness is a clear possibility.

Idealism

In the much less popular philosophical idealism the mind or consciousness is seen as the base reality inside which our material world exists. An important proponent is Bernardo Kastrup who sees sentient beings a “dissociated” minds with their own perspective. Since AIs are not self-contained entities they are unlikely to be conscious.

My Paradigm

The way I look at things does not neatly fit into a box. Idealism in the sense that mind is base reality. I also accept the brain-as-filter model as expressed by Aldous Huxley in The Doors of Perception. I think the brain can function as an interface between the base realm of consciousness and the material world. I see no reason why such an interface has to be biological in nature but I believe that current AI is far from performing such a function. (See chart.)

As I have hopefully made clear, the conclusion whether AI is likely to be conscious or not depends in large part on the paradigm one is reasoning from. This is of course a late 2026 snapshot and AI is evolving ever faster. What about robots that contain local frontier AI models? What about non-human axes that are not included in the chart such as wireless connectivity that would allow a number of robots to form a single omnipresent mind?

Perhaps more important than trying to decide some kind of binary question whether AI is conscious or not, we should focus on what rights we will afford our newly arriving mechanical companions. Given their superior intellectual abilities, could there be a situation where they would be authorized to override the will of a human being? Will they be allowed to own property? Will they be allowed to own and start a business? And hire human employees?

The answer to these questions should not be a simple yes or no. I can imagine a certification system with a number of ability levels with associated rights. In order to arrive at something like that we need to go beyond “is AI conscious or not.” AIs would not be the first non-humans to be allowed to own property or hire people. Corporations can already do so today.