Security firm Raonsecure (042510.KQ) has unveiled an integrated security strategy that gives artificial intelligence agents unique digital identities, much as people carry ID cards, and confines their actions to pre-approved permissions. The approach centers on verifying an AI agent's identity and authority as it makes decisions and carries out tasks, while tracking its activity records to establish accountability.
Lee Soon-hyung, chief executive of Raonsecure, presented the strategy on the 30th at SecuUp 2026, an information security technology conference held at the Fairmont Ambassador in Seoul's Yeouido district. "The more we expand AI's ability to reason, plan and execute, the more the technology that guarantees the safety of those actions must advance as well," the CEO said.
At the heart of the proposal is a "Web3-based AI identity" that embeds the DNA of identity, authority and accountability. Just as a person proves who they are with an ID card, each individual AI agent would be assigned a unique digital identity using Web3 technology. That identity records who delegated which permissions to the agent, and all of its activity is logged on a blockchain.
Raonsecure introduced five solutions to implement the concept: OneAccess, OneShield, OneTag, OneSOAR and OneHacker. Together they link the entire process into a single framework, from the moment an AI agent receives a command through data processing, task execution and threat response. Kim Tae-jin, chief technology officer at Raonsecure, said in a following session that the company did not stop at releasing standalone AI security products. "We designed it by viewing the entire environment in which AI operates as a single pipeline," the CTO said. "What sets us apart is that each technology interlocks organically to protect everything from AI security to AI use."
The event also featured an authentication demonstration using Rawon-i, a humanoid billed as Raonsecure's first AI employee. Rawon-i went through authentication, received a digital employee ID and verified visitors' mobile IDs. In the finals of a hackathon held alongside the conference, 10 teams that advanced to the main round presented ideas built on blockchain and AI technology. "The competition from now on will not be about who builds AI better, but who uses it more safely and effectively," the CEO said. "Raonsecure will combine AI, Web3, security and authentication technologies to build a digital trust infrastructure that people can rely on."
