Until recently, internet users could be divided into two groups: bots and people. Websites have long allowed bots to perform automated tasks such as capturing page content, under rules governing what they can access and how they behave. The web, however, was designed primarily for people. That is now changing rapidly with the release of personal AI agents that can do things for users online, such as booking flights or hotel rooms. Apps like OpenAI's Dots and Meta's Muse have unleashed these agents at scale. Muse was downloaded 5 million times in its first month, and agentic web traffic tripled from August to September. The mixing of agents with human visitors has gone poorly, with many websites blocking AI agents that consumers have sent to negotiate prices or plan vacations.
Agents mark a third category of web visitor. They are not human, and they use the internet at machine speed, but they are not quite bots either, since they act on behalf of individual people to perform legitimate and useful tasks. Because they fit neither category, some platforms, especially security-conscious ones, have banned them outright. Amazon has blocked Muse's agents and is fighting AI startup Perplexity in court. Although a judge ruled in Perplexity's favor, Amazon has appealed, arguing that agents threaten its advertising business by making it harder to tell ads seen by humans from those served to bots.
Ethan Mollick of the Wharton School predicts the blocks will soon affect the whole web. If agents can request refunds or negotiate prices, they could overwhelm customer service departments until they are unusable by real people. Platforms will likely find ways to admit agents, particularly because most agentic traffic goes to e-commerce sites where there is money to be made. Shopify is already giving agents their own checkout lanes that bypass pages designed for humans.
The pressing question is what this means for everyone else. The web may shift toward being optimized for the agents that could eventually outnumber people, leaving humans with worse interfaces and constant verification demands.
Source: MIT Technology Review AI · Summarized by HeadlinesBriefing