Ethereum restaking protocol EigenLayer is introducing artificial intelligence into its decision-making process, seeking to ensure that proposed actions comply with the protocol’s rules.
Working with AI development platform Sentient, EigenLayer will introduce Judge Dobby, which Sentient calls an “AI adjudicator” focused on the subjective decision-making needed to reason through complex, real-world corporate governance scenarios.
“While traditional AI systems excel at tasks with clear, correct answers, such as translation, classification, and recommendations, governance is full of gray areas,” Sentient said in a release. “Current models fail to offer meaningful legitimacy when answering subjective questions, such as who deserves funding in a contested grant round [or] which contributors deserve recognition in an open-source project?”
EigenLayer is the largest Ethereum restaking protocol with nearly $13 billion in total value locked (TVL), according to DeFiLlama. It enables users to restake their staked ETH, using it to provide security for third-party services to earn additional yield. The protocol’s EIGEN token is up 22% in the past month and trades at a $2.7 billion valuation.
EIGEN Price
“Governance in decentralized protocols requires objective assessment of whether proposed actions comply with established rules and community standards,” said Sreeram Kannan, founder and CEO of EigenLayer. “Judge Dobby, built on Sentient’s loyal AI framework… represents an interesting development in the push toward utilizing AI to build more robust and efficient governance for decentralized systems.”
Sentient wants EigenLayer to work with its CHANCERY open-source benchmarking tool, which tests an AI’s ability to perform the kind of complex reasoning required to come up with good solutions in real governance scenarios.
Sentient said its goal is to embed accountability into AI-driven dispute resolution in areas where decision-making is critical, such as corporate governance and public resource allocation.
Read the full article here
