These machines give you a dose of $\mu + \epsilon$ rems, where $\epsilon$ is random with zero mean. TSA's claim amounts to $\mu = 10$ and (presumably) $\epsilon$ normally distributed with $\sigma$ small. Now it turns out that $\mu = 14$. I agree with you that, in isolation, this is inconsequential. Taking a step back though, you have to wonder: if they are incapable of measuring $\mu$ accurately what else have they gotten wrong? In particular, suppose that one time in a million scans this machines messes up and administers a dose of $10000\mu$ (about 1mSv). A few unlucky people a year would be on the receiving end of that -- would you like to be one? This may seem like tinfoil-hat nonsense to you but if you read about the shady way these machines were sold to the government and the farcical "testing" that went on before they were deployed it doesn't seem completely implausible. At least, I'd like to see a lot more openness about how these things work, plus more third-party verification, before really trusting anything coming from the TSA.
But nobody was saying they're wrong except for this author. I question why you're so quick to trust this random author. The author himself explicitly states he emailed the scientists who did this study, and all he says about it is a quick note at the end of this article where they said the amount of error in their results means the TSA could still be telling the truth. IE. The scientists who did this study are saying that their findings aren't enough to prove anything about the TSA, but the author is saying they are. The author also didn't reproduce anything of what the scientists actually said to him. IMO, I don't blame you for not trusting the TSA, but there's nothing about this article that makes it particularly trustworthy either.
