The draft document suggests that the company will place warnings next to Tweets that include manipulated media, and will warn people before they share any such media.
They will also add a link to show why it is thought that the media in question is manipulated.
However, the company has not indicated that it will remove deepfakes when they appear, despite growing pressure on social platforms to deal with the problem.
Twitter has opened the document up to public suggestions, and has said that it will consider responses before finalising its position.