Safe Haskell is a new extension to the Haskell language that is implemented in GHC as of version 7.2. It allows for unsafe code to be securely included into a trusted code base by restricting what features of GHC Haskell code is allowed to access. Put simply, it makes the types of programs trust-able. Safe Haskell itself is aimed to be as minimal as possible to encourage broad adoption by the Haskell community. It provides strong enough guarantees about compiled Haskell code for more advance secure systems to be built on top of Haskell, using techniques such as information flow control security or encrypted computations. These techniques combined with Safe Haskell make Haskell a great language for building reliable, secure multi-party systems today. Particularly relevant with the growing power of web applications and the platform nature of many web sites.
This is work in collaboration with David Mazières, Simon Marlow and Simon Peyton Jones.
From the Haskell Implementors Workshop 2011: http://www.haskell.org/haskellwiki/HaskellImplementorsWorkshop/2011