Business Insider has surfaced an emerging industry argument over whether publishers should serve advertising to AI agents browsing the web on behalf of users. The outlet's reporting frames the question as a contest between competing risks and rewards, with no clear consensus yet among publishers, ad-tech vendors, or advertisers.

The debate hinges on a fast-changing slice of internet traffic. AI agents — automated assistants that fetch pages, summarize content, and complete tasks for users — increasingly request pages the same way human browsers do. Publishers serving ads to those requests generate revenue. But they also risk cannibalizing the human visits that advertisers pay a premium to reach.

Why serve ads to AI agents?

Supporters of serving ads to AI agents point to immediate revenue. A request that returns a page with a display impression is a request that pays, regardless of who, or what, initiated it. For publishers already seeing flat or declining human traffic, agent-driven page loads look like a way to fill the gap.

The pitch from ad-tech vendors tends to emphasize measurability. If an agent identifies itself through standard signals, the request can be categorized, priced, and reported through existing programmatic machinery. That keeps current workflows intact while opening a new inventory pool.

What are the risks?

The risks run in several directions at once:

• Cannibalization of higher-paying impressions. An agent that scrapes a page and feeds the answer to a user may eliminate the reason for a human visit.
• Brand safety. Advertisers paying for human audiences may not want creative running on pages consumed by a bot with an unrelated purpose.
• Ad fraud overlap. Some agent traffic resembles legitimate requests but behaves like fraud, fetching pages it never displays.

Publishers worried about these outcomes want to block AI agents entirely, or at minimum segregate them into a separate, lower-priced inventory pool.

What's the upside for advertisers?

The buy-side case is less obvious. Advertisers have spent two decades building systems tuned to human attention. Most measurement products still count impressions rather than outcomes. An impression delivered to an AI agent may not register in the same brand-lift, click-through, or conversion datasets that justify a campaign's budget.

That asymmetry is at the core of the argument. AI agents consume publisher content at machine speed, sometimes training on it, sometimes answering questions from it. Whether publishers should be paid for that consumption — and at what rate — is the unresolved question Business Insider highlights.

Where the industry stands

No standards body has issued binding rules on agent traffic. Working groups within the digital advertising industry have discussed labeling protocols, but nothing has shipped across the ecosystem. A handful of infrastructure providers now let publishers inspect and filter requests, but adoption is uneven across the open web.

The Business Insider report indicates publishers are splitting into two camps: those treating agent traffic as a new revenue line, and those treating it as a threat to human-monetized inventory. Until a standard emerges, the decision is being made site by site.

What changes if the debate hardens into a rule?

If publishers converge on blocking AI agents, advertisers lose a measurable slice of impressions and publishers gain clarity over what counts as billable. If publishers instead accept agent traffic at par, advertisers may demand price renegotiation or new measurement standards to justify spend against non-human audiences.

Either outcome reshapes the supply side of programmatic. The variable is timing — and how quickly any single dominant publisher or platform moves.