Public key cryptography was a breakthrough because of the many new types of secure communication it suddenly permitted: most importantly, between people who do not have a trusted channel through which to exchage a symmetric key. Instead, it permits each partner to make a public key widely available, as well as use the public keys of others to encrypt messages that only they can decrypt.
One avenue of attack against this kind of system is for an attacker to make a public key available that they pretend belongs to someone else. For instance, you mighy try to impersonate a government or industry figure, then have people send sensitive materials to you inadvertantly. One way to prevent this kind of attack is to use key signing: an approach employed by both the commercial software PGP and the free GPG alternative. With key signing, you produce a web of trust, in which people use their own secret keys to vouch for the validity of public keys posted by others. That way, if I trust Bob and Bob trusts Jim, I can adopt that trust transitively.
GPeerReview is a system intended to extend this trust function to the review of academic work. Reviewers produce comments on documents and sign them with their keys. These comments can include different levels of endorsement for the work being scrutinized.
It is difficult to know whether the level of academic fraud that takes place justifies this sort of cryptographic response, but it seems like a neat idea regardless. Providing secure mechanisms for people to prove who they are and that things are properly attributed to them is increasingly important as technology makes it ever-easier for nefarious individuals to impersonate anyone in front of a wide audience.