Multipartite Einstein-Podolsky-Rosen Steering may lead to ultra-secure internet

"Einstein's reservations about quantum mechanics were encapsulated in the 1935 Einstein, Podolsky, Rosen paradox, which highlighted the theory's strange nonlocality. In this paper, we give theoretical proof that special quantum states allow Einstein's nonlocality to be genuinely shared among many observers, regardless of the spatial separation between them. These states possess a property (which Schrodinger called "steering") that not all quantum entangled states have, and may provide unprecedented security for a future quantum internet.

Suppose Alice has a secret message to transmit. Alice may use the states to send multiple parties a quantum encryption key, or, if the message is highly important, she can "secret share" i.e. distribute the key among the receiving parties, so they must collaborate to decipher the message. The important new feature is that (unlike ordinary entanglement) Einstein's nonlocality cannot be faked by classical means without Alice's collusion. Assuming her station is secure, this enables detection not only of hacking attacks during transmission, but of dishonest receivers or sabotage of devices at receiving stations. We show that the special states can be created in different forms so they are useful for messages carried either as optical amplitudes or as photonic qubits."

http://arxiv.org/abs/1212.2270

And no, this will not lead to ultra-secure internet, which also isn't even mentioned in the article.

1 - It requires direct physical links between the involved parties.

2 - It does not provide authentication, and thus is subject to man in the middle attacks.