Computer Scientists Figure Out How to Prove Lies
10 months ago
- #Cryptography
- #Random Oracle Model
- #Blockchain Security
- Computer scientists have long relied on hash functions to simulate randomness in cryptographic applications, assuming outputs are indistinguishable from true randomness (random oracle model).
- A new paper demonstrates a method to trick proof systems into certifying false statements, challenging the security of the random oracle model, especially in blockchain technologies.
- The Fiat-Shamir transformation, a fundamental technique for verifying computations non-interactively, is vulnerable to attacks exploiting hash function weaknesses.
- Researchers found a way to embed a malicious program in any task, allowing false proofs to pass verification, raising concerns over blockchain and cryptographic security.
- Efforts to patch vulnerabilities in Fiat-Shamir are underway, but the discovery prompts a broader reconsideration of cryptographic assumptions and the random oracle model.