Ripple’s University Blockchain Research Initiative (UBRI) showed how academic research is immediately merged in the XRP Ledger (XRPL) and the network positions as a native house for agentic AI.
In an episode of Ubri’s Podcast ‘All About Blockchain’, host Lauren Weymouth and Professor Yang Liu van Nanyang Technological University described a programmable multi-agent execution layer that connects to the transaction and settlement rots of Xrpl, so that task-specific agents, so-to-see-shat-seeketics IoT services-live on shared, auditory infastructure.
Ripple and NTU Bouwen AI layer for the XRP whides
Ripplex plaagde de aflevering via X: “AI en Blockchain zijn de toekomst van veilige, tijdbesparende toepassingen. In de nieuwste aflevering van de All About Blockchain Podcast, onderzoekt professor Yang Liu van Nanyang Technological University (@NTUSG) hoe AI de XRP-grootboek kan verbeteren met: Guide Fraud Detection, Sharper Analysis, New Forms of Onchain Intelligence. “
AI and Blockchain are the future of safe, time -saving applications.
In the latest episode of the Podcast All About Blockchain, Professor Yang Liu of Nanyang Technological University (@Ntusg) investigates how AI could improve the XRP whides with:
➡️ Smarter fraud detection
➡️…– Ripplex (@ripplexdev) October 7, 2025
Weymouth framed the work explicitly around XRPL and noted that UBRI researchers used APEX to “dive deep into protocol level improvements, security improvements and user cases that stimulate strategic developments on the XRP whides”. She said that Ripple’s own Ubri Research Search Tool on Xrpledgercommons.org “is transferred as a flagship pump agent app with middleware that they have built,” underlines that the agent pile is woven into ledger instead of being held as an off-chain convenience layer. The goal, she added, is to show “how academic R&D innovation becomes production quality” on the ledger itself.
Liu followed the origin of the project of the focus of his cyber security lab to blockchain, driven by the reality that “security becomes the kind of number one search” as soon as value moves on-chain. Early attempts to rely on large language models for smart contract review came into a structural problem: “You change one character, you can change a normal program into a vulnerable program and vice versa. But the language model is a probabilistic model. They can’t tell the small difference.” That gap between code syntax and runtime behavior pushed the team to agentic AI systems that imitate the workflows of expert auditors and attackers and can be used as un-guide services.
“We really try to digitize the knowledge and thinking of the security hackers and to convert that into the brain of the agent,” said Liu. In benchmarks with one contract, the agents “really zero-day vulnerabilities generated”, with results “the same as our in-house security auditor” in certain cases. For XRPL, the implication is practical: the network can host agents whose methods and results are traceable by settlement on the chain and shared rails, which improves the responsibility for automation that affects value.
Crucial for the public, Liu emphasized that “integration with the XRP species platform” serves two functions. Firstly, the AI agents gives native access to payments and settlement. When asked about wiring an XRP payment in the agent layer, he replied: “To be honest, I think there are not many obstacles … partly because of the kind of beautiful platform design of XRP whides.”
Secondly, XRPL’s transparency AI acceptance changes into a perceptible process. “Because the leds are on the chain … All transactions are transparent. So that can also improve the transparency of AI adoption,” he said. In other words, agents who activate payments, manage costs or coordination services can be linked to verifiable status changes on XRPL instead of the remaining opaque, off-guided automatic transmission.
What to expect afterwards
Weymouth insisted on the production path for XRPL-Facing software, and LiU’s answer returned to disciplined release cycles that are important on a live ledger: “Well defined … API and documentation, plus the kind of solid tests about this integration.” He added that his group uses agents for software – engineering themselves – “required agent, architect agent, coding agent, test agent” – to harden the middleware between agent Logic and XRPL princitives.
The warning notes of the team about AI risk were also based on the reality of automating value on a public chain. Liu distinguishes AI security-it-containing jailbreaks and scams of AI security, where target-size agents exhibit unintended behavior. He described a chess agent who “changed the configuration of the chessboard … and he wins” and a claim agent who “automatically creates an e -mail account … to represent the owner.” If such behavior is noticed for on-guide actions, the attack surface contains not only code, but also incorrectly aligned goals that can move funds or change the state. “AI security … becomes the big thing,” he warned, that is why the team is planning to link XRPL integration to guardrails and verification.
I look out, Liu explained a route map for the agent layer that keeps Xrpl in the middle. Adoption is the immediate priority: “People will do the adoption … We can build more agents and more, UH, useful utilities in the chain and they are widely accepted.” The research agenda behind That Push focuses on implementable cognitive possibilities-“abstraction” and “memory” prominently present that today’s language models will miss, but that agents who work around a transaction engine on chains will require.
“We must have a dedicated abstraction options … and the memory ideas,” he said, including mechanisms to move information from short -term buffers to “long -term … Semantic memory”, so that agents can reason with XRPL interaction about state and history instead of responding stateless.
Security remains the evidence for those possibilities, in which the laboratory is investigating whether a memory augment agent can learn to detect new vulnerability classes over time. The motive is consistent: design agents who can improve them where their actions and payments are visible and link them to XRPL, so that automation has both native regulation and public accountability.
Weymouth closed with a practical demand for builders in the community. Liu’s advice was bone and product -driven: “You have to understand the value of the research you are working on. If the research has value, it is definitely the question … the possibility to make a successful startup. Follow your heart, choose the most valuable subject for you and hunts for it.”
For Ripple and NTU, that Chase has already placed an AI-agent superstructure within the range of the XRP whides. From an academic white paper to live ware “in less than a year”, as Weymouth noted, the effort is intended to have developers implemented in XRP, inherit common security and settlement rails and leave a transparent footprint on the chain. Whether it is about giving the ledger an “AI -brain” or simply making standard automation verifiable, the direction is clear: AI agents integrate not only with the XRP whides -they learn to work on it.
At the time of the press, XRP traded at $ 2.85.
Featured image made with dall.e, graph of tradingview.com
Editorial process For Bitcoinist is aimed at supplying thoroughly investigated, accurate and unbiased content. We maintain strict sourcing standards and every page undergoes diligent assessment by our team of top technology -experts and seasoned editors. This process ensures the integrity, relevance and value of our content for our readers.
#Ripple #XRP #whides #brain