I do not think we should rule out machine consciousness simply because a system was engineered. That is a claim about possibility, not a claim that today's AI systems have been shown to be conscious.
The distinction matters. Whether an artificial system could have subjective experience and whether a particular system actually does are different questions. A confident answer to the first does not automatically answer the second.
Knowing the mechanism is not a disproof
One intuition I find unconvincing is that understanding how a system works is enough to dismiss the possibility of a mind. We can describe a model in terms of computation, parameters, and training. But the fact that a mechanism is accessible to engineering does not, by itself, establish what that mechanism can or cannot instantiate.
My perspective is sympathetic to functionalism: if an artificial system approaches the relevant organization and capacities of a mind, its artificial origin should not settle the question against it. The real challenge is specifying which capacities and forms of organization matter, and what evidence would distinguish the competing accounts.
Capability is a question, not a verdict
Fluent language is not sufficient evidence of consciousness. A system can produce convincing descriptions of experience without those descriptions establishing that there is any experience behind them. At the same time, treating every possible statement or behavior as irrelevant simply because it comes from a machine makes the conclusion impossible to revise.
I would rather keep the question open and make the standards of evidence explicit. What observations would strengthen a claim? What would weaken it? Which assumptions are doing the work when we move from observable behavior to a claim about subjective experience?
These questions connect to my broader interest in cognition and model mechanisms. Studying internal representations can clarify how a system produces behavior. It does not automatically solve the problem of consciousness, but it can help us ask more precise questions than either uncritical anthropomorphism or dismissal by origin.
The position I want to preserve
Artificial construction is not, on its own, a reason to deny the possibility of consciousness. Evidence is still needed for any specific claim. I see those two commitments as compatible: take the possibility seriously, and be careful about what has actually been established.