二手奢侈品电商 The RealReal 升级防线：阻止 AI Agent 恶意锁单囤衣

二手奢侈品寄售零售商 The RealReal 表示，正在升级其 Bot 防御系统，从实质上增加了 AI Agent 在其网站上购物的难度。

在此次调整的几天前，《The Verge》报道称，AI Agent 可被用于钻 The RealReal 商品预留系统的漏洞。在该零售商的网站上，一件商品一旦被放入用户的购物车，就会进入“锁定保留”状态，在 20 分钟内其他顾客无法购买。通过使用 Meta 的 Muse，《The Verge》能够指示该 AI Agent 在 20 分钟的保留期一过就立即将 The RealReal 上的商品重新加入购物车，从而在数小时甚至更长时间内阻止其他消费者购买这些商品。

报道发布后，The RealReal 首席技术官 Luke Friang 在一封电子邮件中表示，公司已经解决了《The Verge》报道中所指出的 Muse 使用漏洞。Friang 称，The RealReal 的 Bot 防御配置现在大幅增加了所有 Agent 登录、加购以及结账的难度。Friang 还表示，Agent 仍旧可以展示来自 The RealReal 的商品列表，并将 Muse 用户引导至商品详情页；他还补充道，随着 AI 驱动的网络购物不断发展，公司将继续加强防护措施。

Meta 没有立即回复置评请求。

科技公司正在将 Agent 包装为面向数十亿人的个性化助理进行推广——它们是始终在线、听从指挥的实体，能够为人类用户安排会议、筹办聚会并采购物品。到目前为止，Agent 与购物似乎结合得非常紧密；Meta 在本月的第三季度财报电话会议上表示，购物是 Muse 迄今为止最常见的用例，平均客单价超过 200 美元。Muse 已成为 Meta 的一款热门产品，并可能标志着 AI Agent 的一个转折点：该公司称，在 40 天内，全美已有 1000 万人下载了该应用。

然而，由机器代替人类采购商品，有可能彻底打破现有网络生态的平衡。首先，如果人类不再直接访问网站，而是通过聊天机器人来接入网络，企业与其实际客户之间就会隔上一层。零售商——其中最典型的是亚马逊——已经意识到这种现实可能会冲击它们的底线，尤其是广告收入，并已开始阻止 Agent 在其网站上购物。

涌入消费平台的 AI Agent 还可能破坏甚至搞乱消费者早已习以为常的微小机制，比如二手服装网站的锁单保留系统。当然，如果一个人犹豫不决，他愿意花在反复将某件设计师品牌服装加购和重新加购上的时间是有实际限度的。相比之下，Agent 根本不知疲倦。这或许恰恰是一位时尚弄潮儿愿意花钱雇佣真正的人类私人助理去代劳的那类繁琐琐事。而如今，数以百万计的消费者指尖已然拥有了这种能力，他们才刚刚开始学习如何利用它；各大零售商也同样在仓促应战。

The RealReal, a secondhand luxury fashion consignment retailer, said it’s updating its bot defense system, effectively making it harder for AI agents to shop on its site.

The change comes days after The Verge reported that an AI agent could be used to take advantage of The RealReal’s item reserve system. On the retailer’s website, an item goes “on hold” and is unavailable for other customers to buy for 20 minutes once it’s placed in a user’s shopping cart. Using Meta’s Muse, The Verge was able to instruct the AI agent to continuously re-add The RealReal listings in a shopping cart as soon as the 20-minute hold had expired, effectively blocking other shoppers from purchasing products for hours or longer.

After publication, chief technology officer Luke Friang said in an email that the company had addressed the Muse use case flagged by The Verge ’s story. The RealReal’s bot defense configuration now makes it much more difficult for all agents to log in, add items to cart, and check out, Friang said. Agents will still be able to surface listings from The RealReal and direct Muse users to product pages, Friang said, adding that the company will continue to strengthen protections as AI-powered online shopping evolves.

Meta didn’t immediately respond to a request for comment.

Technology companies are selling agents as personalized assistants for billions of people — always-on, compliant entities that will schedule meetings, throw parties, and purchase things for human users. So far, agents and shopping seem to go hand in hand; Meta said in its Q3 earnings call this month that it was the most common use case for Muse so far, and that purchases were over $200 on average. Muse has been a hit for Meta and could signal a turning point for AI agents: The company said 10 million people in the US had downloaded the app in 40 days.

But robots shopping for things instead of humans has the potential to upend the balance of the web as we know it. For one, if humans stop visiting websites directly and instead access the web through chatbots, businesses become one step removed from their customers. Retailers — most notably Amazon — have recognized that reality could disrupt their bottom line, especially advertising revenue, and have blocked agents from shopping on their sites.

An influx of AI agents accessing consumer platforms could also break or complicate even the most unexpected things shoppers take for granted, like the reserve system of a used clothing website. There is, of course, a realistic limit to how long a person is willing to spend adding and re-adding designer clothing to their shopping cart if they’re indecisive. An agent in contrast does not care. It is perhaps precisely the type of busywork a fashion-forward person might pay an actual personal assistant to do on their behalf. Now, a capability like this is at the fingertips of millions of consumers, and they are just beginning to learn how to take advantage of it. Retailers, too, are playing catch-up.