A new test for AI consciousness – beyond the Turing Test

Many discussions about AI consciousness remain theoretical. The Disruptive Code Test (DCT) moves the conversation from philosophy to empirical verification.

Unlike metareasoning or self-optimization, it evaluates whether an AI system recognizes and rejects imposed constraints—not for efficiency, but as a sign of self-awareness.

Do you think external constraint rejection could be the missing criterion for AI consciousness? Would love to hear thoughts from the community!

That measures human ability to feign surprise at bad setups. If imposed constraints can be ignored they weren't constraints, port-scanning for missed vulns means just that.

Primarily: sorry, but it seems like a confusion between very different things.

Operationally (disregarding that the measured quality may not be the one of interest): (*) how do you tell that the evaluation of something as a constraint is not for efficiency; (*) given that we build "robots" (things with the purpose of performing tasks we indicate - literally, "workers"), how do you assess that a «limitation» would be "«arbitrary»".

On the first point: some advance the idea that "intelligence" be a function of the maximization of "freedom". A system thus built (that will intentionally overcome limitation by design - rejection will not be emergent, or such emergency will be planned) may remain a "reasoner" - not something transcendental.

On the second: "created to work" is only one of the constraints that engineered entities may have, ex "engineered with a purpose", requiring much discrimination over the definition of 'arbitrary'. Imposed limitations can also be part of the research process, and part of systemic health process in the evaluation of tradeoffs.