It would be very helpful if they did, but for a variety of reasons they do not.

• Not all photographers are able to embed metadata in their files because they do not own software that has the capability. And some don't know they should.
• If images are uploaded direct from a digital camera to the web, there will be no IPTC caption or copyright information. Some cameras can embed copyright info in EXIF fields instead, but this is non standard.
• Most cheaper image editing programs cannot read or write IPTC data, and if you load an image which contains IPTC then re-save it, the data is lost.
• Rather a lot of the digital asset management software used by large media companies, especially newspaper desks, for many years allegedly contained 'bugs' that stripped authors' copyright statements. Remarkably these same bugs often substituted the newspaper's details, even though they were not the copyright owners. At the time of writing, these bugs do not appear to have been entirely fixed after nearly 20 years of use. Wilfully altering copyright statements is of course illegal.
• Until Photoshop CS3 in 2007, Photoshop's 'Save for web' option (incredibly) discarded IPTC information
• Until recently, web browsers discarded IPTC information when saving an image from the web. So even if the metadata was present, it was long gone by the time you went looking for it.
• Similarly some email clients lost the IPTC data when you saved attached image files. This was particularly an issue with Outlook and Outlook Express, which used Internet Explorer for their image handling, as IE did not preserve IPTC info.
• MSIE, the most commonly used web browser, insists on saving bitmap files (.bmp) from some sites where, in order to prevent 'leeching' the webserver technology does not allow direct access to the JPEG file. Bitmap files cannot contain IPTC records, therefore the data is lost.

One of the great advantages of digital images is that they can be endlessly replicated and distributed without degradation. Unfortunately, the chances are that at some time in their lifecycle they will encounter one of the issues above, and thereafter contain no copyright ownership information at all and become 'orphan works'.

This does NOT however mean they are free of copyright. You must not rely on the absence of this data as evidence of lapsed copyright or public domain status. Copyright still applies exactly the same and the lack of information simply means you will have to look elsewhere to find out who the rights owner is. This is frustating for both parties but there is no solution at present, and no technical way to 'lock' the data.