Even if humans were significantly prone to Hyper Active Agency Dection (H.A.D.D.), a large study showed that this plays no role in the intuitive appeal or wide-acceptance of theism.
University of British Columbia study (2014): “Anthropomorphism, operationalized as the tendency to project human-like attributes to non-human entities, was not related to belief in God in our model. In our adult sample, it was not related to belief in God even in a zero-order correlation. This may be surprising given theories that argue that anthropomorphism and hyperactive agency detection are an underlying feature of all supernatural belief (Barrett, 2000, 2004, 2008; Guthrie, 1993, 1996) … Given influential theories that place the origin of religious belief in anthropomorphism, it was surprising that the path from anthropomorphism to belief in God was non-significant. One might wonder whether this null result is a reflection of any problems with the IDAQ – a particular measure of anthropomorphism ( Waytz, Morewedge, et al., 2010 ). Given that this is a validated scale with good predictive power, we find this unlikely. However, in order to rule out this possibility, we fit this model a second time using the alternative, task-based visual measure of anthropomorphism (Norenzayan et al., 2008), with a moderate-to-high correlation with the IDAQ r (490) = .47, p < .001. We found similar fit results that confirmed the previous findings with the IDAQ (Yuan–Bentler X 2 (4, N = 492) = 6.19, p = .19; CFI = .99; RMSEA = .03) (see Fig. 2), suggesting that the null finding regarding anthropomorphism is not an artifact of the particular measure we used.” [Aiyana K. Willard, Ara Norenzayan “Cognitive biases explain religious belief, paranormal belief, and belief in life’s purpose.” Cognition 129 (2013): 379-391.]
Intuition is generally reliable. (That is to say, although there are plenty of cases where it misfires, it is rationally taken to be trustworthy for cases where there is an absence of some reason to think it is misfiring. They are analogous to our perceptual faculties, and should not be dismissed simply because illusions exist.)
But so what? That's because reliable intuitions are more adaptive. Mutations leading to reliable intuitions get passed on more effectively to offspring. God-belief evolved simply because hyper-active agency detection is adaptive, not because it is true.[See response]1
Helen De Cruz: “…sometimes agency detection does go awry, as when we hear wooden planks creak in an old house and form the belief that there's a burglar in the house. But many more times, we form the belief that there is an agent, when there actually is an agent (e.g., when you see someone walking across the street from you on a clear day. When debunkers of religious belief appeal to hyperactive agency detection, they are already assuming that the agent that is being detected (e.g., God) is of the false-positive kind, the sound in the dark room. But I don't see how they can assume this in a non-question begging sense.” [“Thoughts on Evolutionary Debunking Arguments Against Religious Belief” (Sept 2012): Online.]
The idea that humans evolved through an unguided evolutionary process (e.g. Darwinian evolution) is false. [Forthcoming] This is relevant because if unguided evolution is false, then theism's intuitiveness could not have been an unguided accident.
God does not exist. This is relevant because it implies that the theistic inclination to believe in God or gods exist must be “hyperactive” (i.e. misleading and erroneous). Also, given Darwinism, our tendency towards agency detection must have been “selected for” simply on grounds that it was adaptive for survival; the truth or falsity of theism is irrelevant.
It would be adaptive for humans to be oversensitive to detecting agency in the world.
But so what? Many things that would be adaptive for humans do not end up evolving. There is no reason to expect then that this would have evolved.