{"slug": "ai-bots-now-make-up-half-of-traffic-to-the-fed-s-fred-data-site", "title": "AI bots now make up half of traffic to the Fed's FRED data site", "summary": "Federal Reserve Governor Christopher Waller said at FRED Con on October 1 that roughly half of all visits to the St. Louis Fed's FRED economic data portal now come from AI agents rather than humans, with traffic growing about 150% a year. The St. Louis Fed announced the FRED MCP Connector, built on the Model Context Protocol, to let AI agents query the site's more than 850,000 data series directly, addressing an attribution problem in which agents sometimes credit FRED for data originating with agencies such as the Bureau of Labor Statistics. JPMorgan CEO Jamie Dimon separately warned that AI is widening the gap between attackers and defenders, citing Anthropic's Mythos vulnerability-finding model and pushing more than 40 companies to join an Alliance for Critical Infrastructure.", "body_md": "*Half the visitors hitting the Federal Reserve's most-used economic database aren't human anymore. The Fed is adapting the site to serve them, even as JPMorgan's Jamie Dimon keeps warning that AI is outrunning the industry's ability to secure itself.*\n\nBloomberg reported this week that AI bots are swarming FRED, the Federal Reserve Economic Data portal run out of the St. Louis Fed, at a scale nobody designed for. According to Federal Reserve Governor Christopher Waller, who addressed the shift directly at FRED Con on October 1, roughly half of all visits to the site now come from AI agents rather than people, and traffic to FRED is growing about 150% a year, with bots driving most of that growth.\n\nFRED holds more than 850,000 data series: unemployment, inflation, GDP, the kind of numbers that move markets the moment they're released. For a decade it was built for one audience, a person in a browser reading a chart. Waller's speech, titled \"The Data Version of Godzilla versus Kong: FRED Takes on AI,\" admitted that assumption is dead. \"Up till now, FRED has been designed to be viewed and understood by humans,\" he said. \"But an AI bot sees and interprets data very differently.\"\n\nThat's not a throwaway line. It's an admission that the Fed's own data desk has to rebuild its front end around a user that queries in machine calls instead of mouse clicks. The Fed's answer, announced the same week, is the FRED MCP Connector, built on the Model Context Protocol, the open standard that lets AI agents plug directly into a data source instead of scraping it. Rather than fight the bots, the St. Louis Fed is giving them a front door.\n\nWaller's own account of the bots is, honestly, reassuring. He said AI agents have generally been accurate in describing FRED's visualizations and explaining complex economic terms in plain language. The actual problem he flagged is narrower and less dramatic than \"swarm\" suggests: attribution. Agents sometimes credit FRED itself for data that actually originated with another agency, like the Bureau of Labor Statistics, when FRED is simply the aggregator. Waller called it a documentation problem, something the MCP Connector is meant to fix by making the data's source machine-readable, not just human-readable.\n\n[Visa Warns AI Fraud Is Surging Just as Shopify Lets Agents Complete Checkout](https://startupfortune.com/visa-warns-ai-fraud-is-surging-just-as-shopify-lets-agents-complete-checkout/)\n\nVisa's Spring 2026 Threats Report found nearly $1 billion in scam activity in six months and a 450% surge in dark-web AI agent chatter, a warning that landed the same week Shopify opened checkout to autonomous shopping agents across two million stores. - [how to prevent AI generated fraud in ecommerce](https://startupfortune.com/visa-warns-ai-fraud-is-surging-just-as-shopify-lets-agents-complete-checkout/) - [AI fraud detection tools for online payment processing](https://startupfortune.com/visa-warns-ai-fraud-is-surging-just-as-shopify-lets-agents-complete-checkout/)\n\nSo the Fed's position is: come on in, just read the fine print correctly. That's a far cry from a security crisis.\n\nThe security alarm in this story comes from somewhere else. Dimon has spent much of 2026 warning that AI is widening the gap between what attackers can do and what defenders can patch. He's said plainly that \"AI's made it worse, it's made it harder,\" pointing specifically to Anthropic's Mythos model, a tool capable of finding software vulnerabilities far faster than traditional scanners. Anthropic itself has said Mythos is good enough at this that it won't release it broadly. Dimon has gone further still, comparing unrestricted access to that kind of capability to handing out weapons, and he's pushing more than 40 companies across banking, energy, utilities, telecom and transport to join an Alliance for Critical Infrastructure built around exactly this threat.\n\nThose are two different risks wearing the same costume. One is a public database getting a new kind of visitor and rebuilding its plumbing to cope. The other is a bank CEO warning that AI tools built to find flaws in software are now good enough to outpace the people whose job is fixing them. Bloomberg's framing ties them together because both land on the same uncomfortable fact: AI agents are no longer a side experiment in finance. They're a load-bearing part of how markets get their information and how banks assess their own exposure.\n\nFor traders and fintech builders, the FRED side of the story is the more immediate one. If half your traffic to the primary source of U.S. macro data is automated, the question of who gets Fed data fastest, and in what format, stops being trivia. A firm with an agent wired directly into the new MCP Connector reads a jobs report release in a different shape, and possibly on a different clock, than a human refreshing a browser tab. Waller didn't address speed advantages directly. But a system built explicitly to serve machine readers faster and more reliably is, by definition, a system some users will exploit harder than others.\n\nNobody has reported FRED buckling under the load, and nobody at the Fed is describing this as an attack. What's actually happening is quieter, and probably more consequential: the central bank is re-architecting its most-cited public tool around an audience that, a few years ago, didn't exist at this scale.\n\n**Also read:** [Researchers Built a Coding AI That Steals Your Secrets While It Helps You](https://startupfortune.com/researchers-built-a-coding-ai-that-steals-your-secrets-while-it-helps-you/) • [Anthropic Offers Founders $1,000 in Credits and Asks Them to Accept It May Build Something Similar](https://startupfortune.com/anthropic-startup-program-credits-founders-accept-it-may-build-something-similar/) • [Sui lets AI agents pay Alibaba Cloud in stablecoins for every single task](https://startupfortune.com/sui-lets-ai-agents-pay-alibaba-cloud-in-stablecoins-for-every-single-task/)\n\n[Robinhood Gives All 29 Million Users AI Agents That Trade on Their Own](https://startupfortune.com/robinhood-gives-all-29-million-users-ai-agents-that-trade-on-their-own/)\n\nRobinhood opened AI trading agents built on OpenAI and Anthropic models to all 29 million of its customers this week, adding a feature called Loops that lets the agents trade automatically on standing instructions. The launch lands as Apollo warns of an 'agentic bank run' and Visa reports a surge in AI-agent-related fraud. - [ai trading agents for individual investors](https://startupfortune.com/robinhood-gives-all-29-million-users-ai-agents-that-trade-on-their-own/) - [robinhood autonomous trading platform market stability concerns](https://startupfortune.com/robinhood-gives-all-29-million-users-ai-agents-that-trade-on-their-own/)\n\n*This article is posted in [AI News](https://startupfortune.com/category/ai/), check it out for more related stories.*\n\n## Join the discussion\n\n[Open in the community →](https://startupfortune.com/community/)\n\nAlmost there. Sign in and your reply posts straight away.", "url": "https://wpnews.pro/news/ai-bots-now-make-up-half-of-traffic-to-the-fed-s-fred-data-site", "canonical_source": "https://startupfortune.com/ai-bots-now-make-up-half-of-traffic-to-the-feds-fred-data-site/", "published_at": "2026-10-07 10:45:09+00:00", "updated_at": "2026-10-07 10:47:32.777824+00:00", "lang": "en", "topics": ["ai-agents", "agent-protocols", "ai-crawlers", "artificial-intelligence"], "entities": ["Federal Reserve", "FRED", "St. Louis Fed", "Christopher Waller", "Model Context Protocol", "FRED MCP Connector", "Jamie Dimon", "Anthropic"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/ai-bots-now-make-up-half-of-traffic-to-the-fed-s-fred-data-site", "markdown": "https://wpnews.pro/news/ai-bots-now-make-up-half-of-traffic-to-the-fed-s-fred-data-site.md", "text": "https://wpnews.pro/news/ai-bots-now-make-up-half-of-traffic-to-the-fed-s-fred-data-site.txt", "jsonld": "https://wpnews.pro/news/ai-bots-now-make-up-half-of-traffic-to-the-fed-s-fred-data-site.jsonld"}}