Ex Machina remains one of the most philosophically astute films addressing AI, agency, and consciousness. Upon rewatching, the depth of its exploration of AI manipulation seems prescient, echoing current discourse surrounding AI's potential for subtle psychological control and alignment risks.
However, the film makes one conceptual slip: it conflates the Turing Test—a measure of conversational indistinguishability—with a consciousness test. The original Turing Test, proposed by Alan Turing, explicitly avoids addressing consciousness. Instead, it pragmatically defines intelligence as behavioral indistinguishability in conversation. Passing the test means merely that we have lost any objective grounds, based solely on behavior, to deny intelligence or thinking to the tested agent.
But what does "passing" actually signify? Turing intentionally sidestepped subjective concepts like consciousness or genuine understanding, recognizing these as philosophically unresolvable. Instead, passing implies that intelligence is observable, measurable, and not uniquely human. It challenges our assumptions about what thought and intelligence truly entail.
Interestingly, ChatGPT itself passes the Turing Test easily, raising questions about whether the original test was too narrow or easy. Historically, AI repeatedly moved the "goalposts": first Chess, then Go, now conversational indistinguishability. Each advancement revealed that these benchmarks alone did not capture the deeper, richer qualities humans intuitively associate with true intelligence and agency—qualities like intentionality, causal reasoning, and strategic planning.
This continuous repositioning of the goalposts isn't merely arbitrary shifting but necessary conceptual refinement. Each time AI surpasses a benchmark, it clarifies what we genuinely value and aim to measure.
An intriguing philosophical intuition emerges from this process: perhaps consciousness itself might be an inevitable byproduct of sufficiently advanced agency. Genuine agency—requiring robust internal modeling, strategic planning, and self-reflection—naturally implies the development of subjective awareness as a necessary "interface" for integrating internal states and external interactions. Thus, consciousness may not be an optional add-on but an emergent side-effect of sophisticated agency.
Ultimately, Ex Machina’s conflation of the Turing Test with consciousness inadvertently illustrates the very conceptual challenge we face today: refining what it truly means to be intelligent, agentic, and possibly conscious in the age of advanced AI.