A first step towards automatic hoax detection

Recent hoaxes, specially virus-related hoaxes, have shown a previously unknown degree of danger. Although current policies against hoaxes are certainly helpful, their ability for dealing with them is very limited, especially against some of the new and more imaginative types of hoaxes outlined in this article. The expansion of the Internet, and the corresponding boost in hoax evolution, have increased the need for automatic detection tools, and everything suggests this need is going to grow in the near future. In this work, after a quick introduction to the past, present and future of hoaxes, their quick change rate and the difficulties and limits for their automatic detection are highlighted. The paper proposes two approaches: heuristics, which should be enough for dealing with the easiest and most common ones, and traffic analysis to fight the toughest.