Drawbacks of Web Robots

They use bandwidth as they surf the Web and retrieve sites
WebBots can stress/crash servers with their "rapid fire" (requesting many pages from the same server in a short period of time) retrieval of pages
Only known pages are harvested: WebBots work by knowing of one server and "jumping" to servers linked to by the known server. This leaves an estimated 5% of Web pages isolated, as they are not linked to other pages
WebBots don't do their work in real time: it takes a day or three from the point where the WebBot discovers the page to when that page can be retrieved by a search engine working in concert with the WebBot.
WebBots can be turned away from sites
(see the following page, The Standard for Robot Exclusion for details)
rogue bots are programmed to purposely crash servers by continuously requesting large numbers of pages; WebBots can also be programmed to retrieve copies of pages or entire sites for loading elsewhere as the work of someone other than the true author
WebBots can be programmed to cast votes in polls, etc. (one recently threw off a Web poll done by People Magazine)
hits made by WebBots can inflate page hit statistics: when a WebBot hits a site, the hit rate of the site will spike; a Webmaster may not realize those hits were an infrequent WebBot hit and not a trend. Advertisers, learning this, are sometimes reluctant to depend on page hit counts as the costing criteria for Web ads.