Database Makes Theory

A New Digital Turn on the Horizon

In late July of last year, Facebook shut down its artificial intelligence robots after they appeared to be developing language in which to speak to each other.  One of the robots had uttered: ‘i can i i everything else… you i i i everything else…’[1] bizarrely echoing an excerpt from Faulkner’s The Sound and the Fury: ‘i i will never do that nobody knows what i know and he i think youd better go…’

Moving past the awkward shift into some existential state mirroring Quentin Compson’s neurotic self-awareness, we come to the more practical matter of this new digital turn.  It is worth discussing since the last digital turn, the one we still seem to be wrapping our heads around, shook up the world of history so profoundly. What changes are in store this time around?

All our talk thus far has been centered on the fact that computers cannot interpret the data, and that it is still up to us to continue to play with it and make conclusions of our own.  With the introduction of artificial intelligence, however, the new charge that ‘the database is the theory’ may evolve further into ‘the database makes the theories.’

Seem too much like science fiction?  Need I remind you that AI itself was once considered science fiction? If AI can diagnose a medical problem, why not a social upheaval, given its light-speed analysis of all primary (once digitized) and secondary source content? AI can now make music and art, forms of expression that were once considered unique to the human condition, and what does this say of the interpretive abilities of artificial intelligence?

The Proposal

The print project I propose is to investigate this matter. AI is already being programmed to recognize objects and events, and before you ask, “how can it be considered AI if it has to be programmed?” consider that children too must be programmed to recognize their environment to some degree. Once recognition is hard-wired the real learning can take place, and at some point, we should address the fact that this will lead to interpretive, actionable reasoning on the part of the computer.  Whether or not this can be carried over into the abstract is a matter of opinion, or perhaps ‘administrator permission.’ Either way, should any of this become a reality, what then does it mean for the discipline?

There is plenty of room to speculate.  What would it even mean if AI could produce interpretive work?  Doesn’t the nature of interpretation mean that it is always up for debate? Perhaps, but suppose AI presents some airtight logistical analysis of human behavior and tendencies, and can think up data based evidence that no human being could even consider formulating. A computer might know how to manipulate the data in ways humans could never dream of, and it is on this point that AI could process new applications for not just history but any field.  It is likely that once the machine learns to learn, it will far surpass human intellect, but what role does emotion play in the interpretive process? Will AI ever be able to acquire that human capacity? And would it even want to?

[1] http://www.independent.co.uk/life-style/gadgets-and-tech/news/facebook-artificial-intelligence-ai-chatbot-new-language-research-openai-google-a7869706.html