In fact most images on the Internet are not copyrighted and are royalty free. The 'reality' of the situation (best practice) is that you can use images at will, and if someone does come knocking then you take the image off.
How to spot copyright images? For example, Reuters is famous for using Getty images, and images on their website are clearly marked as having originated from Getty -this is a big no no then, but to be honest, people still take them and use them freely. Again, unless someone comes knocking at the door and you ignore them then you are not in trouble.
It aint so pretty, but it is the reality
Future developments. Reverse image searches. Google haven't done it yet, but someone else has and their index is growing. TinEye,
http://www.tineye.com, for example will accept an image from yourself and search it's index of billions of URLs for uses of that image, and is intelligent enough to recognise crops, blurs etc etc. The index is not all encompassing as Google's but when Google does buy them (haha) the 'reverse image search' technology will be a copyrighter's dream tool.
I am not offering legal advice, but quite simply the reality. Even the BBC do it, I have seen the Guardian do it. Until someone comes knocking, nothing to worry about; just plead ignorance and comply when they do.
Obviously, don't attempt to copyrighte the image yourself of course or incorporate it into any copyrighted material such as a logo.
Majestic~