I joined X/Twitter recently in an effort to be noticed by the senpais within the field of consciousness studies. In general, I’m not big on social media but I also recognize that it can be useful for reaching a larger audience. It’s interesting to see what people are talking about, especially with respect to AI.
One tweet that has recently generated a lot of discussion is one published by Pope Leo XIV on artificial intelligence:
He is correct, but of course, his claims are informed by a Catholic point of view. Personally, I am agnostic so though this tweet resonates with me, it does so for different reasons than it might with others.
AIs do not undergo experiences because they are not built in a way which is analogous to living beings. Even if they were to possess a body, that doesn’t mean that they have experiences. The reason is because experiences rely on semantic information which computers do not have access to. Robots like iCub may look like they experience elements of the environment, but they do not because their bodies are not built in an analogous manner to biological creatures. They do not possess affect and motivation, and lack the perception/action loops identifiable in organisms. It isn’t enough to have a body to produce experience, it’s the functional organization inherent to that body which matters.
The Pope is correct that they “do not feel joy or pain” because they aren’t built to have these signals impact their behaviour in the relevant ways. The if-then contingencies that constitute behaviour are insufficient to be classified as affect because these operations are descriptions of the real deal. There is no integrated system of affect, perception, and action; the code AIs and robots run on are abstractions of the physiology of biological systems.
AIs do not know what anything means because they don’t work with meaning, they work with rules. If a threshold has been met or crossed, subsequent functions are executed as a result. They cannot judge good or evil because this requires an evaluation of qualitative information. Though qualities can be represented by quantitative information to a degree, there will be aspects that cannot be fully captured. The colour red is not reducible to frequencies and wavelengths, it depends on which cones and colour columns are being activated in the retina and brain of the viewer.
“They may imitate or even simulate, but they do not understand what they produce, for they lack the affective, relational, and spiritual perspective through which human beings grow in wisdom.” The first part of this sentence is correct, but a bit more explanation is required when we get to “relational and spiritual perspective.” I’m not exactly sure what he means by ‘relational’ though I think I can guess, and the Pope’s appeal to spirituality may not resonate with everyone. The appeal to wisdom is more poetic than a necessity, because understanding can be achieved regardless of whether wisdom is developed as a result. It might be, but I would leave wisdom out of it. We know that the Pope is interested in exalting humans given that we are supposedly made in the image of God, so this claim is understandable regardless of its truth value. For the most part, humans do grow in their ability to make wise decisions and offer sound advice to others as they move through life, but this isn’t a necessity. Humans are magnificent, but I find most living beings magnificent in their own way, regardless of the metaphysics behind their existence. AIs and robots are interesting and potentially useful, but my concern is equating their behaviours and capacities to those of biological organisms. They are artifacts created by humans, and we must keep this in mind as their abilities progress and expand over time.
Neural Blender has stopped working for me, even after creating an account. It just hangs and says my image is being generated ad infinitum. As such, Neural Blender and I are no longer friends. I have a new friend: https://aifreeforever.com