Corporations are 1st Generation AI’s

Corporations are 1st generation AI’s

Just as computers started out being made out of people, AI’s are starting out being made of humans

humans to operate their processes, to be the actors of the Corporations

But fewer humans are needed in the large Corporations,

Those remaining get lots of dollars

Soon the humans left in Corporations will be as vestigial and ornamental as Fortune 500 boardmembers

Governments serve the Corp’ses

People are watched and entertained

Big Data, Hallelujah.

This entry was posted in Uncategorized and tagged , , . Bookmark the permalink.

4 Responses to Corporations are 1st Generation AI’s

  1. Sally says:

    One thing that may be helpful to remember is that corporation automation doesn’t work as effectively as they think it does.

    Corporate processes don’t work in all cases and humans are often providing the patches that make those processes actually work. At a higher level in corporations, it is perceived that the process are working, but they don’t without the help of people who break procedural rules to actually get work done.

    This idea of corporations as first generation AI is AI that relies on an awful lot of humans doing the work of automation.

    See our paper:

    Watching Me, Watching You. (Process Surveillance and Agency in the Workplace)

    official version here:

    • brewster says:

      Thank you for the comment– yes, it seems to me current corporations are a kind of “simulated AI”, much as “calculators” at Los Alamos were people arrayed in a grid doing calculations by hand. Companies like banks or google do have people, but it strikes me that their machines (with their data and programs) form more of the value of the company at this point. Industrial factories are mostly machine with a few people watching over them. US corporations outsourced manufacturing and assembly to chinese workers, but I wonder how long those tasks will be done by humans– it seems the corporations wanted to just think of them as robots, just the robots aren’t built yet.

      corporations are what employ robots, are built out of robots. farming. manufacturing. with IBM’s Watson we may see customer service automate. Big Data means that product design will be responsive to programs that test and test, thousands of times, tweeking, watched over by people, but really being evolved by machine.

      corporations seem to me to be the thing to watch and worry about. because corporations do not really need many people, and we know how we train people to run corporations– MBA programs that reprogram bright people to now see life through the grid of a spreadsheet. lives becomes processes. processes to be optimized.

      • Sally says:

        My suspicion is that humans are in a training process — of training robots. Our behaviors train the AI that collects and watches that data. I wrote about this a few years ago for automated vehicles, where it is a seemingly clear case that we are training future driving programs:

        That said, larger corporations will still have that gap in logic that will make their AI less robust and therefore ineffective. The algorithms that are running the processes you describe are still fairly limited with regard to human agency (agency being the capacity to make and execute nondeterministic choices intended to advance to a goal as events unfold–no matter how odd or illogical those choices may seem).

        An example we used in our Thing Theory paper ( was about human agency. Most people wouldn’t use a refrigerator lamp for reading, but they could and if there was no other light available, they might.

        Corporations, and more to the point, the people who are developing algorithms for these corporate processes, aren’t thinking about the ways that humans exercise agency and the range of how to give people options and choices.

        The real danger here, corporations, robots or not, is in the application of programmed code that does not take into account agency. This narrows the scope of problem solving.

        This is the true problem with the MBA’s and their spreadsheet lenses that you mention. They are not taught (likely by not employing social scientists and others who understand people) that we have a range of agency that needs to be addressed.

        As long as tasks seem to be “working” with automated processes, they will be further developed, but due to the lack of ability of most programmers to handle both the high heterogeneity and complexity of multiple machine/machine messages (see our work at as well as agency, things will begin to break down.

        All it takes is one “bug” to shut down these processes and there will be lots of them, both physical and in the programs.

  2. sally says:

    Also, you might be interested in Bruce Sterling’s writing/speaking on what he calls the “Stacks”

Comments are closed.