Improving the internet for AI agents could make it worse for everyone else

Meta description: Personal AI agents from OpenAI Dots and Meta Muse are forcing websites to rewrite rules for bots, humans, and automated shoppers as agentic traffic triples.

Tags: AI agents, Meta Muse, OpenAI Dots, web automation, Shopify checkout

Featured image: abstract technology scene with glowing network paths, agent icons, and checkout lanes, no real people, no company logos, no text

AI agents are becoming a third category of web visitor, and the recent launch of apps such as OpenAI Dots and Meta Muse has pushed websites to rethink rules written for people and bots. The shift matters because automated visitors acting for individual humans could reshape shopping, customer service, and the design of the web itself.

From a Binary Web to a Third Visitor

For years, internet traffic could be sorted into two groups: bots and people. Websites let bots perform automated tasks such as capturing page content, while rules governed what they could access and how they behaved. The web was designed primarily for people, even when people had to solve puzzles and captchas to prove they were not bots.

Personal AI agents now break that split. They are not humans, and they use the internet at machine speed. They are also not quite bots, because they act on behalf of individual people to perform tasks that can be legitimate and useful. They simply complete those tasks faster than a person would.

The Recent Surge and Its Early Fallout

That third category arrived at scale with the launch of apps such as OpenAI Dots and Meta Muse. Muse was downloaded 5 million times in its first month, and agentic web traffic tripled from August to September. These agents can do things for users online, including booking flights or hotel rooms.

The mix of agents and human visitors on a web built for people has gone poorly so far. Many websites are blocking AI agents that consumers send to negotiate prices or plan vacations. Still, if demand for agents keeps rising, the internet may be redesigned for them instead of for people.

Because agents do not fit existing systems for separating bots from humans, some platforms, especially security-conscious ones, have chosen to ban them entirely. TechCrunch recently reported that people had trouble deploying their agents on websites ranging from Pizza Hut to Yelp, which does not allow AI to book restaurant reservations.

Amazon has blocked Muse agents and is in a court battle with AI startup Perplexity over its agents as well. A judge recently ruled in favor of Perplexity, but Amazon filed an appeal arguing that AI agents threaten its advertising business by making it harder to distinguish between ads seen by humans and ads served to bots.

Ethan Mollick, an associate professor at the Wharton School of the University of Pennsylvania, predicts the friction will soon affect the whole web. Once it becomes common knowledge that AI agents can be sent to request refunds or negotiate prices, agents could overwhelm customer service departments to the point where they become unusable by real people. He recently wrote on LinkedIn, “Humans will eventually get very annoyed when their agents aren’t allowed to do this on their behalf.”

E-commerce Opens Dedicated Lanes

If demand for personal agents grows, platforms may find ways to let them in. That is especially likely because the vast majority of agentic web traffic goes to e-commerce sites, where allowing agents to shop for real people may create revenue. Shopify is already welcoming agents by giving them dedicated checkout lanes that bypass pages designed for humans.

The question is what happens to the rest of the internet. People have grown used to a web that often exploits human attention, but humans have remained the intended users. A future optimized for agents could mean neglected sites for real people, worse browsing interfaces, imperfect verification systems, and constant prompts to prove one is not an agent. Prices could also vary by the agent requesting them, and recommendations could be shaped by advertising or platform commissions in ways users cannot see.

Rules Still Being Written

That outcome is different from the future promised by Google, Anthropic, and OpenAI, where AI agents give consumers greater power and bargaining leverage. The promise is not outlandish: agents could help fight bogus medical bills or research more affordable car insurance.

Whether they can do so depends on the rules set for how agents interact with websites and what they are allowed to do. Those rules are still being written. Some will benefit everyone, because neither consumers nor businesses want credit card information to be waved around by roaming agents, and stronger agent security is welcome. The contested choices are whether businesses can block agents or stop them from negotiating better deals. Those decisions will determine whether the web becomes a highway for agents working for people, or against them.