There is a specific genre of content that I’ve been receiving lately. (I think emphasizing the passive is important here.) Whatever governing forces calculate and decide what my eyes should see, as one small slip of short-form media follows the next through the squinting frame of my hand…well, those forces have brought me an aggregate of AI-critical content.
To be more specific, my phone has been showing me clips of men critiquing AI-generated outputs. In one series, the conceit is that a man is directing AI actors as if to produce a commercial product. In these videos, he (a real person) gives specific directions to the AI actors, who immediately fail in some surreal way. The scenarios can be quite varied. The AI actors attempt to load a dishwasher. The AI actors are trying to perform a double entendre. The human director attempts to correct them, gives more specific directions. They continue to fail. The failures escalate. In another, singular video–two well-known podcasters play clips of radio programs put together by various AI. The results are incongruous, and, once more, absurd. One of the AIs reportedly refuses to make the program because, “there’s no audience that needs this.” The two podcast guys are laughing a lot. The conceit with the director is that he’s in a professional scenario, but he does laugh as well–and the comment section under each of his videos is exploding with laughter.
Some people leave comments tensely questioning the logic of using (indeed, paying for) AI software in order to critique artificial intelligence. Like all satire, the risk of ironic affirmation is high. I’m not sure we can really call it an unintentional irony. In each instance, there is also a certain subtext: AI will not replace us,…yet. In each instance, we get a sense of how critique and complicity are two facets of the same object.
The explosiveness of the laughter in response to AI is striking to me. The AI generations themselves are, of course, inherently ugly. Grotesque, in fact. Grotesque in ways that don’t even have to be explained–and that’s, ultimately, what the laughter is about. There is something involuntary about this forceful reaction to something so stupid, so wrong, and so unnecessary. But there’s also, in turn, something ugly about the laughter itself. In part, the laughter comes from a recognition that the results of the generations are such abject failures. Yet the inquiring subjects are still implicated in the failures under examination. Derision is itself a human technology, one which creates and affirms the boundaries of what can and cannot be accepted. These AI outputs are derisible precisely because they dance around that boundary of acceptance, creating a mirror image that is simultaneously recognizable and unrecognizable. One of the reasons why these AI failures are “funny” is because the scripts which run behind the output are searching with an insincere and incomplete form of volition. And in the end, what is derisible is the concept of automated volition.
