Coinbase AI sends false World Cup score before match start

Coinbase’s AI‑generated news system sent thousands of users a push notification on Sunday claiming Norway had defeated Brazil 3‑2 in a World Cup knockout match—hours before the game began—prompting immediate scrutiny of the exchange’s AI‑driven alerts as it pushes deeper into AI‑powered prediction markets【1】.

AI misfire and the underlying system

The erroneous notification appeared on Sunday, displaying a fabricated final score and crediting Erling Haaland with two goals—details that later proved partially correct when Norway actually won 2‑1 and Haaland did score twice. Coinbase’s head of consumer and business products, Max Branzburg, confirmed the mistake and said the company “fixed the incorrect story and made some updates to avoid these types of inaccuracies in the future”【1】. The alert was generated by Coinbase’s AI news engine, which pulls data from its sports prediction market on Kalshi, a federally regulated platform for binary‑style wagers. Because the AI treats high‑probability market prices as factual outcomes, it can mistakenly broadcast speculative results as confirmed news【3】.

Business context and incentives

Coinbase has been expanding beyond pure crypto trading, rolling out AI‑powered wallets, agents, and a suite of prediction‑market contracts covering sports, elections and economic data. The firm launched its Kalshi integration in January, allowing users to trade event contracts tied to real‑world outcomes【1】. Analysts estimate the prediction‑market line generates roughly $100 million in annual run‑rate revenue, while Kalshi invested $20 million in World Cup advertising alone, underscoring the financial stakes driving aggressive user alerts【3】. Earlier in the year, Coinbase also faced criticism for a March‑Madness targeting bug, for which CEO Brian Armstrong publicly apologized, yet the AI‑driven alerts continued to surface【3】.

What to watch

• Future AI alerts – Monitor Coinbase communications for any further AI‑generated notifications that could misrepresent real‑time events.
• Prediction‑market activity – Watch trading volume and price swings on Kalshi contracts for high‑impact events, as sharp moves may trigger premature AI alerts.
• Regulatory response – Track any statements from U.S. regulators or consumer‑protection bodies regarding the use of AI in financial‑service notifications.

The incident highlights the tension between rapid AI‑driven user engagement and the need for reliable, verified information. As Coinbase leans further into AI and prediction markets, the accuracy of its automated alerts will become a litmus test for trust in the platform’s expanding product suite.