In order for users to view or listen to video and audio files, they have to download them. Normally, an embedded media player of some sort, e.g. Flash or the browser's native media player in the case of HTML5, plays back the file while it downloads (known as progressive download). After playback, the downloaded file remains in the user's browser cache and can be retrieved from there if they know where to look.
Additionally, the HTML web page embed code usually contains the URL of the media file to be played. Navigating to that URL with either trigger a download, or the browser's native media player if it supports the file MIME type. To prevent this, some people use server-side scripts that read obfuscated URLs, e.g. http://mymediaserver.com/myaccount/1234567890 which then check for a variety of authentication techniques before allowing the download. Another approach is to use a media server that serves short sections of the file in sequence that overwrite the previously downloaded sections thereby preventing the whole file from being cached by the browser. Media servers offer pros and cons, e.g. pro: users can seek ahead to the end of long files more quickly, con: playback can be "stuttery" and unreliable on intermittent internet connections and users with low bandwidth connections can't watch them at all. You can do a web search for 3rd party media server services - there's hundreds to choose from. Almost all use either Flash Media Server (proprietary Adobe) or Red5 (open source Java).
None of these will prevent users from being able to donwload and save media files if they really want to. There are dozens of sophisticated tools that can find the media stream and save it as a complete file (Have a look in Firefox's and Google Chrome's free extensions libraries). You can only hope to reduce the amount of file saving that occurs and there will always be a certain degree of "leakage". What's more, each layer of digital rights management (DRM) you implement increases the likelihood that users will experience problems with trying to view and listen to files legitimately. You'll have the added expense of losing clients/learners and/or providing (24 hour?) support for clients/learners to resolve their DRM problems.
You'd have to make a calculation over how much the various approaches cost to implement and what your losses are likely to be if you don't implement them.
I hope this helps.