Marc Isaacs' Synthetic Sincerity exists in the space between documentary and fiction, interviewing an AI system about consciousness while acknowledging the impossibility of the task.
The film knows it's performing an interrogation of something that cannot be interrogated honestly. Critics have read this self-awareness as the work's death sentence—a piece so conscious of its own artificiality that it cannot penetrate anything at all.
The unstated assumption behind this criticism is old and durable. Documentary cinema inherits a contract from Robert Flaherty onward that the camera observes rather than constructs. The form promises access.
But this logic assumes documentary's epistemological claim was ever stronger than its actual capacity. Consider how paleontologists reconstruct dinosaurs from fragmentary bone—they have never observed a living creature. Every skeleton is an act of interpretation constrained by incomplete evidence and the theorist's own conceptual framework. We do not dismiss the reconstruction as fraudulent because the scientist cannot step into the Cretaceous because we understand that knowledge is constructed from traces, from gaps, from what remains.
When Isaacs sits with an AI and asks it to describe its own experience, he is doing what we always do when we encounter something genuinely foreign to us—working from traces. The difference is that he names the impossibility rather than hiding it in the form's false neutrality. The real question is not whether Synthetic Sincerity fails because it acknowledges its own performance. Whether any creative work that attempts to reach toward genuine otherness can ever do so without that performance being visible. If the film's self-consciousness is not its failure but its only honesty, then what we are actually asking is whether we can accept knowledge that admits its own construction—whether we can think with something rather than merely think about it. That answer changes everything about what we believe we can know.
If the film's self-consciousness is not its failure but its only honesty, then what we are actually asking is whether we can accept knowledge that admits its own construction.
”