Duplicate content
Duplicate content is the same or nearly the same text showing up at more than one web address, on your own site or across sites.
Duplicate content is the same, or nearly the same, text appearing at more than one address. It can happen inside your own site, when one page is reachable at several URLs, or across sites, when someone republishes your article or you reuse a supplier's product descriptions.
It's rarely a penalty. Most of it happens by accident, and search engines just pick one version to show. The trouble is that they may not pick the one you want, and the credit for links and visits gets split between copies.
Where it comes from
Most duplicate content is technical. The same page with and without www, with tracking codes added to the address, in a print version, or listed under two categories in a shop. Each address looks like a separate page to a crawler.
The rest comes from reuse: product text copied from the manufacturer, press releases posted word for word on many sites, or articles republished by partners.
Copying other people's content in bulk to fill a site is a different matter. Google's spam policies treat scraped content as spam.
How to fix it
Pick one address for each page and point everything else at it. A permanent redirect is the strongest way to do that, and a canonical tag on each copy is nearly as strong. Google explains both in its guide to consolidating duplicate URLs.
For example, Harrow & Pine's oak dining table can be reached at four addresses because it sits in three categories and appears in a sale. Adding a canonical that points at /tables/oak-dining-table tells search engines which one to show and gathers the links onto it.
Duplicate content and AI answers
AI answers cite pages by their address too. If one page lives at four addresses, citations can be spread across all of them, and the page looks less used than it is. When your text is copied onto other sites, an answer may also cite the copy instead of your original.
In AnswerOrbit
Sources lists every page the answers used, by its address. If the same page of yours shows up twice under different addresses, it needs a canonical or a redirect. See Sources.