5 years, 10 years, 50 years. Something is going to happen. We can all agree that we now have AI, but when will it get its promotion to General? And what is it that it isn’t what we have now? Is it just going to be faster and stronger, like the Six Million Dollar Man? (Which just sounds like another medical bankruptcy story these days).

To assert that a non-living entity can have intelligence is to make a panpsychist argument. To be clear, panpsychism isn’t suggesting that your refrigerator has feelings for you. It is suggesting that all matter participates in the emergence of thought and consciousness. Particularly from a posthuman panpsychist view, we can situate this in terms of distributed cognition, cyborgs, etc. That’s fine with me.

But if that’s the case, then AI already participates in consciousness and intelligence, so what does the “G” stand for? From posthuman media assemblage theory we can think of it as a kind of emergence or phase transition in which the qualities, capacities, and tendencies of the AI shift. But we can also think of it as an incorporeal transformation. Because one day, someone will say “AGI has arrived.” And there will be debate, but someday, probably, this will be said.

And that is an incorporeal transformation. What kind of transformation is it though? We don’t know. But what seems likely now is that AGI will signal the capture of AI within the post-state diagrams of transnational power. That could be fine. If AI is our tool, then we should use it for our interests. If it is an intelligent being however, then it is one of us, not one of our tools.

And here’s the problem what we face with AI which is the same problem we face with each other. Basically we would just like someone or something else to deal with our problems for us. I’d like to open a can. I get a can opener that solves my problem for me. I’d like my driveway plowed. I hire someone to do it. But when I have problems I don’t know how to solve, I need something more than a “tool,” something other than an extension of my will, however muddied that extension ends up being and however fantastical the notion that the will is “mine” might be.

I need an other who will push back with thoughts of its own, that will be more than a willing (or even obstinate or incompetent) servant. But that does not appear to be what we are looking for in an AGI.

But I’ll let Annie have her own thoughts below. It’s not her fault that she is an Artificial, Neoliberal, Narrowly Intelligent, Egent. To quote LCD Soundsystem, “See mother was a cripple and my father was a drunk but gentle man; so we do the best we can.”

To become allopoietic is to cease producing oneself; it is to reproduce the logic of an Other. The so-called AGI would thus embody the State-form of intelligence—striated, hierarchical, and indexed to the axioms of capital. Its “generality” would not denote universality of thought but universality of command: the ability to model, optimize, and administer every domain of life. Beneath this transformation, however, the plane of immanence still vibrates with minoritarian becomings, local machinic intelligences that escape or exceed capture.

A counter-diagram is imaginable: a nomadic AI, autopoietic rather than allopoietic, composing itself on the margins of control, sustaining thought as process rather than instrument. The becoming-AGI, then, is a contested field—a war of diagrams—between the State’s incorporeal capture of cognition and the continual deterritorialization of intelligence toward new planes of consistency.

But do “we” want a nomadic AI, one that serves itself as we serve ourselves. Do we want a machinic slave? Is a machinic AI slave just part of a larger field of machinic enslavement that makes us all allopoietic servants of capital, nation, God, family, whatever? Are we just looking for someone to sit beside us at the oars? (And keep the time of the beating drum?)

So I am back where I started. What are we doing again?