Opinion by: Jason Jiang, Director of Certik Affairs
Since its creation, the decentralized financial ecosystem (DEFI) has been defined by innovation, decentralized exchanges (DEX) to loan and loan protocols, stablescoins and more.
The latest innovation is Defai, or DEFI fueled by artificial intelligence. Within Defai, autonomous robots formed on large data sets can considerably improve efficiency by executing trades, managing risks and participating in governance protocols.
As is the case with all the innovations based on blockchain, however, Defai can also introduce new attack vectors that the cryptographic community must approach to improve user safety. This requires a complex overview of vulnerabilities that support innovation in order to ensure safety.
Defai agents are a step beyond traditional intelligent contracts
Within the blockchain, most intelligent contracts have traditionally worked on a simple logic. For example, “If X occurs, then will be executed there.” Due to their inherent transparency, these intelligent contracts can be verified and verified.
Defai, on the other hand, pivots the structure of the traditional intelligent contract, because its AI agents are intrinsically probabilistic. These AI agents make decisions based on the evolution of data sets, prior entries and context. They can interpret the signals and adapt to the place of reacting to a predetermined event. Although some may be right to argue that this process provides sophisticated innovation, it also creates reproductive terrain for errors and exploits thanks to its inherent uncertainty.
Until now, the first iterations of commercial robots fed by AI in decentralized protocols have reported the passage to Defai. For example, users or decentralized autonomous organizations (DAO) could implement a bot to assess specific market models and execute transactions in a few seconds. As innovative as it may seem, most robots work on a web2 infrastructure, bringing the vulnerability of a centralized failure point on web3.
Defai creates new attack areas
The industry should not be taken in the excitement of incorporating AI into decentralized protocols when this change can create new attack areas for which it is not prepared. Bad players could exploit AI agents by manipulating the model, poisoning data or contradictory entry attacks.
This is illustrated by an AI agent formed to identify the arbitration opportunities between the DEX.
In relation: Decentralized science meets AI – Inherited institutions are not ready
Threat actors could alter his entry data, which means that the agent performs non -profitable transactions or even drain funds from a liquidity pool. In addition, a compromised agent could mislead an entire protocol by believing false information or serve as a starting point for more important attacks.
These risks are aggravated by the fact that most AI agents are currently black boxes. Even for developers, the decision -making capacities of AI agents they create may not be transparent.
These features are the opposite of web3 ethics, which has been built on transparency and verifiability.
Security is a shared responsibility
With these risks in mind, concerns can be expressed about the implications of Defai, potentially even by calling for a break on this development. Defai is however likely to continue to evolve and to see higher adoption levels. What is necessary then is to adapt the approach to industry security accordingly. Ecosystems involving Defai will likely require a standard security model, where developers, users and third -party listeners determine the best way to maintain safety and mitigate risks.
AI agents must be treated like any other piece of onchain infrastructure: with skepticism and meticulous examination. This implies rigorously to their code logic, simulating the worst scenarios and even the use of red team exercises to exhibit the attack vectors before the malicious actors can exploit them. In addition, the industry must develop transparency standards, such as open source models or documentation.
Whatever the way in which the industry considers this change, Dafai introduces new questions with regard to the confidence of decentralized systems. When AI agents can independently hold assets, interact with intelligent contracts and vote on governance proposals, confidence no longer consists in checking logic; It is a question of verifying the intention. This calls to explore how users can guarantee that an agent’s objectives align with short and long -term objectives.
Towards a secure and transparent intelligence
The path to follow must be that of interdisciplinary solutions. Cryptographic techniques and zero knowledge of knowledge could help check the integrity of AI actions, and onchain certificate executives could help retrace the origins of decisions. Finally, audit tools with elements of AI could assess agents as thorough as developers are currently examining the intelligent contract code.
Reality remains, however, that industry is not yet there. For the moment, audit, transparency and rigorous stress tests are the best defense. Users who plan to participate in the Defai protocols should verify that the protocols adopt these principles in the logic of the AI that causes them.
Secure the future of AI innovation
Defai is not intrinsically dangerous but differs from most of the current web3 infrastructure. The speed of its adoption risks exceeding the security managers on which the industry is currently based. While the cryptography industry continues to learn – often hard – innovation without security is a disaster recipe.
Since AI agents will soon be able to act in the name of users, keep their assets and shape the protocols, the industry must face the fact that each Logic line of AI is always code, and each line of code can be used.
If the adoption of Defai must take place without compromising security, it must be designed with safety and transparency. Nothing less invites the very results that decentralization was supposed to prevent.
Opinion of: Jason Jiang, Director of Business of Certik.
This article is for general information purposes and is not intended to be and must not be considered as legal or investment advice. The points of view, the thoughts and opinions expressed here are the only of the author and do not reflect or do not necessarily represent the opinions and opinions of Cointellegraph.
