Don Boudreaux has a nice summary of the problems with a lot of product safety problems.
You write as if "safe" is an objectively determinable and unique fact, such as whether or not your newspaper's paid circulation exceeds 500,000 or whether or not your sister is pregnant. But "safe" is not objective in this way. Because no product is 100 percent certain never to cause even the slightest harm (or 100 percent certain to cause harm), the question "Is this product safe?" has no correct single answer. It has correct answers as varied as the number of that product's potential users. No product is "safe" or "unsafe" in the abstract.
Perhaps your tolerance for risk is higher than mine. Perhaps the pleasure I get from using a product is less than yours. If so, should I be permitted to prevent you from using that product because, for me, the product is insufficiently safe? My evaluation of the product's safety is correct only for me, not for you. And matters don't change if I'm a government official.