Vitalik Buterin examines ways to reduce bribery threats.

A team of researchers from Cornell University is investigating the potential risks of a “dark” voting system in Decentralized Autonomous Organizations (DAOs).

This team was founded by Ethereum co-founder Vitalik Buterin and PhD students Mahimna Kelkar, Kushal Babel, Philips Daian and James Austgen. Their work revolves around how to reduce the risk of decentralization when DAOs go mainstream: integrated attacks on protocols with the bribery of smart contracts.

At the Blockchain Science Conference held at Columbia University in early August, Cointelegraph spoke with Mahimna Kelkar about the team's research on Proofs of Complete Knowledge (CK) – a new cryptographic concept they introduced in 2023.

Proof of knowledge is a cryptographic concept that allows one party (the verifier) ​​to let another party (the verifier) ​​know some secret information, such as a secret key, without actually revealing the information.

The concept has been widely used in the crypto industry to improve the privacy of transactions, but a “hidden loophole” still allows situations where this secret information can be carried out by some external method, such as trusted hardware, rather than the prover directly. According to Kelkar:

“When the secret key is stored in trusted hardware, in what we call encryption encryption, you can still complete this knowledge verification without knowing the secret key.”

Bribery attack

This limitation on the way formal knowledge proofs are defined makes voting protocols vulnerable to bribery attacks, Kelkar explained.

The lack of a central authority is the key concept behind the governance of DAOs. DAO members are usually token holders who have the power to vote on rules and decisions. In a bribery attack, however, a malicious actor can offer financial incentives to people through smart contracts, bribing participants to vote for a particular proposal or outcome.

“[..] The voting platform may be vulnerable to bribery attacks. […]A place where users sell their votes for bribes in the dark market,” Kelkar explained. “What our job is trying to do is establish an individual, real human kind of data ownership.”

Proof of complete knowledge

An attacker can use a Trusted Execution Environment (TEE) to ensure that token holders who receive bribes cannot vote freely. In this environment, the attacker controls when and how keys can be used.

The researchers identified two ways to ensure complete knowledge. One involves using a TEE to ensure that a voter has a key and can use it. The token holder can remove the key of this area for free use at any time.

This way, tokens still have full control over their keys. Even if an attacker wants to lock the key to control the voter, the key is managed by the TEE of the voting system itself.

The second approach involves limiting keys using application-specific integrated circuits (ASICS), which are the machines that are often used in Bitcoin mining. By sending the key to the ASIC – without the TEE environment – the key remains accessible to the user, ensuring that they have full control over it, while still ensuring that the key is used by the ASIC and preventing its use in the TEE.

According to Kelkar, the research is still in the prototype stage. “We show that this is a real threat to DAOs, and we demonstrate this by demonstrating a dark DAO that can be implemented in practice, which facilitates voice buying in an existing DAO. It's not something you can deploy tomorrow, but as a research example today, it's practically fast,” added Kelkar.

Magazine: DeFi and Ethereum Are ‘The New Narrative': Michael Van de Pop, X Flame Hall