Would you care to show any non-ICR/AiG references for this?
Yes, that's the usual creationist mantra. The trouble is, radioactive dating is based on some extremely simple physics. We can measure the decay of 14C, or 238U, or 40K in a laboratory, and we know it in every case obeys exact first order kinetics, with a half-life we can measure to high accuracy. If we take 40K, for example, we can measure the quantity of that isotope in a rock, compared to stable 39K. And we can measure the amount of 40Ar trapped in the rock. We know that rock doesn't normally contain a lot of inert gas, and even if it does, we can measure the amount of stable 39Ar and correct for it. So, by looking at the decrease in 40K and the increase in 40Ar, and using simple decay kinetics, we can get an age.
Occasionally, it is true, we run into problems. The most usual problem is if some or the 40Ar has escaped from the rock (if the rock has been heated at some point this often happens.). However, as dating methods have become more sophisticated, we run into such problems less and less. And in any case, if there is a problem with the sample, it means we get an age of 60 million years rather than 120 million years; we don't get an age of 4000 years!
No measurement scheme is perfect. But unfortunately for you, "flawed" isn't good enough to rescue creationist dogma. You need to be able to say something like, "radioactive dating methods are totally non-functional", but you can't. Radioactive dating methods consistently give very ancient ages for rock layers all over the Earth. Occasionally an answer is demonstrated to be substantially off, and when it is it creates a big problem, as in the famous KBS tuff controversy. In that case, a 1.9 million-year-old layer was dated to be 2.4 million years old. The discrepancy turned out to be caused by an inadequate sample collection method. It was eventually corrected--such is the nature of science--but in as precise a science as paleontology, an error of 25% can be a very big problem. In this case, it led to inconsistencies in the fossil record that were spotted in very short order.
This example should give everyone an idea of just how sensitive a scientific probe the fossil record can be. Even inconsistencies of a fraction of a million years can be detected. Such is the level to which evolution can be tested observationally.
No they are not, but you're welcome to present your argument as to why you think they are.
You are further invited to explain why:
1. Independent (repeat, independent) dating methods which rely on no common assumption give the same date results for the same samples.
2. Dating methods give correct dates for items of known age.