The answer is in their robots.txt. http://www.metallica.com/robots.txt has
This means that we have crawled zero pages from www.metallica.com. If you look carefully, we show a description of www.metallica.com from the Open Directory Project, not any snippet from their page. SafeSearch can’t return uncrawled/empty documents (unless they have been whitelisted), because the documents might turn out to be unsafe. Hopefully it makes sense that SafeSearch shouldn’t return a document to a user if we don’t know what that document actually has in it–what if the document had porn on it? So while we could whitelist metallica.com, the correct answer is for their webmaster to allow us to crawl their site.
It’s easy to see how this misunderstanding could happen. For example, if you do the search [Nissan Motors] you get back a pretty useful snippet: “Manufactures automobiles including passenger cars, buses, trucks and related parts and accessories. (Nasdaq: NSANY).” It almost looks like we’ve crawled the page–but we haven’t. Nissan also forbids all search engines from crawling their site with a robots.txt, so that snippet also comes from the Open Directory Project.
Several years ago, the Library of Congress had a robots.txt that didn’t allow any search engines (they do now), so it www.loc.gov wouldn’t show up in SafeSearch. So we changed it so a whitelist can trump an uncrawled document in SafeSearch. We studied the .gov domain and didn’t find any pornographic content (the closest we found was the Kenneth Starr report).