(English Version: Top-right translate button OR scroll down.)
在签名前，用户往往难以直接理解交易后果，因此AI可以通过模拟交易执行结果、预览交易的后果，帮助用户理解这次交易将导致什么资产变化、权限变化、合约调用，从而使用户在签名前可以识别潜在的风险、降低信息不对称。
在提交交易后、链上执行前中，由于调用路径复杂、调用层级太多，用户可能难以理解这些行为对应交易里的哪些具体后果，因此可以借助AI帮助识别高风险或者异常交易。在部分场景下，系统可以根据预设策略对高风险交易进行拦截、标记、告警、要求进一步确认或自动拒绝。
交易执行结束后，状态变化对于人类来说未必是直观可读的，因此可以借助AI翻译成直观可理解的语言，转译成资产、权限、风险或业务结果，用户便能知道交易带来了什么结果；
在此基础上，AI可以在预设边界内替代部分高频、规则明确的人工响应。AI现在最适合可被提前定义的风险判断与响应，而涉及高成本、责任重的交易操作，还是需要人类介入参与决策。
Before signing, users often find it difficult to directly understand the consequences of a transaction. Therefore, AI can help users understand what asset changes, permission changes, and contract calls this transaction will lead to by simulating transaction execution results and previewing transaction consequences. This allows users to identify potential risks and reduce information asymmetry before signing.
After submitting a transaction and before/during on-chain execution, users may find it difficult to understand which specific consequences in the transaction these behaviors correspond to due to complex call paths and too many hierarchy levels. Therefore, AI can be used to help identify high-risk or abnormal transactions. In some scenarios, the system can intercept, mark, alert, require further confirmation, or automatically reject high-risk transactions based on preset strategies.
After the transaction execution is completed, the state changes are not necessarily intuitively readable for humans. Therefore, AI can be used to translate them into intuitively understandable language, converting them into assets, permissions, risks, or business results, so that users can know what outcomes the transaction has brought. Building on this, AI can replace some high-frequency, rule-clear manual responses within preset boundaries. AI is currently best suited for pre-defined risk judgment and response, while transaction operations involving high costs and heavy responsibilities still require human intervention in decision-making.