The rush to AI is built on top of an idea that digitalisation and data can replicate and comprehend being human. The big datasets and models now coming with personalities and fluent conversations.
Yet this smooth completeness is superficial. It ignores the subtleties that actually make up human experiences.
Embodiment matters
There are two ways that AI fails now (and will always fail).
- Qualia
- Durée
Both are centred around the simple fact that humans are embodied and AI is not
Humans are born and live their lives in bodies in places and communities.
AI are coded and exist, when activated, within data centres.
Embodiment is so obvious it is ignored. It is why good research is founded in recognition of living experiences and the social fabrics and contexts of being alive as a human. Bad research tries to break apart a human from their embodiment. Controls and measures are put in to remove ‘bias’. Yet the bias that research removes is the humanity of the participants.
Qualia matters
Qualia matter because they are the experiences of being embodied as a human. The sensory and emotional perceptions made from a perspective that is held by being in a body over time and thru experiences of life and communities.
This is the subtlety of human experiences in moments that data cannot record and AI cannot comprehend.
Metaphor, poetry and storytelling attempt to communicate qualia. Yet we all know that they are private and personal. We can recognise how people perceive and take pleasure in divergent ways but, as books like Language versus Reality, it is impossible to exactly define and record such experiences. If we could clearly define and describe qualia as quantitative data we would. But we cannot and no amount of data centres will change that.
The qualia are in the bodies we live in. They are in the moments and places of experience.
They are in our being.
Durée matters
Durée matters because humans experience their lives as a continuous life in time and place.
We are born, we grow and we encounter new things, new people and new experiences. All of these stretch backwards thru our memories of time and place.
There is a continuous thread in an embodied life. We may often do the same things or go to them same places but the experiences are not the same because we have a lived experience as well as a living experience.
AI cannot have this for two reasons.
Firstly, as noted in the Qualia section, the data and digitalisation process are either not subtle enough to notice the experiences or, more problematically, the digital snaps the thread. All those bits of data fracturing the analogue continuity. The digital tries to replicate the real but never can as it will always be fragments and chunks.
Secondly, the AI do not exist continuously. They are code that activates when requested. They are active when requested by prompts. They do not live lives over time. They exist on demand. This is convenient for businesses but does not replicate human experiences.
Humans matter
All of this is to say that any human, anywhere, of any age, is more interesting and startling more capable than any AI.
The brilliance of AI is only there if you deliberately ignore what humans are and ignore how much AI borrow from humans in prompts.
This is also not a fixable problem for AI developers. They cannot make Qualia and Durée. They cannot make an AI with living and lived experiences of places, people, time and communities.
AI has its uses but it is, and always will be, less deeply capable than whoever the next person you meet is.
Rather like Human Rights, there is a question about the Human Capacities.
The universality of human rights is found in breaking the idea that some people have more rights because of ancestral achievements, physical prowess or property ownership. People have rights because they are people and and an equitable society ensures that.
The universality of human capacities is found in breaking the idea that people are only meaningful in the capacity to work. This is the modern trap of framing life within work. Education and assessment all as categorising and certifying capacities that are relevant to task completion.
AI is more capable than humans if your framing is work. Humans are more than work and their capacities are much vaster. Qualia and Durée are human capacities and they are the subtleties that show the edges of capacities that data and AI cannot comprehend.