Coinbase AI sends false World Cup alert before kickoff, CEO investigates
In brief
- Coinbase AI sent false breaking news alert claiming Norway beat Brazil 3-2 in World Cup match at 10:26 a.m. ET, before 4 p.m. kickoff
- Alert falsely stated Erling Haaland scored twice; actual result was Norway winning 2-1 with Haaland scoring twice
- CEO Brian Armstrong investigating the issue with his team
- Coinbase made updates to prevent future AI-generated inaccuracies
The Alert and the Backlash
Coinbase sent users a false breaking news alert claiming Norway's soccer team beat Brazil 3-2 in a World Cup knockout match before the game started. The alert stated that Erling Haaland scored twice in the match at MetLife Stadium. The timing made the error obvious: the alert arrived at 10:26 a.m. ET, while the match didn't start until 4 p.m.
The actual match result was Norway beating Brazil 2-1, with Haaland scoring twice—so the final scoreline was wrong, even though one detail about Haaland's performance was correct.
Company Response
Coinbase CEO Brian Armstrong replied to a post saying he was looking into the issue with the team. Max Branzburg, Coinbase's head of consumer & business products, stated the incorrect story was fixed and the firm made updates to avoid similar inaccuracies in the future.
"It's awesome to see the power of AI-enabled 24/7 insights for trading, but obviously still need to tune it to address these types of issues." — Max Branzburg, Coinbase head of consumer & business products
Prediction Markets and the Broader Push
Coinbase rolled out prediction markets to U.S. users in January through Kalshi, offering contracts tied to sports, elections, economic data and other real-world events. The company has framed prediction markets as part of a broader effort to become an "everything exchange."
Coinbase added stock options, pre-IPO markets and an AI adviser as it tries to fold more financial activity into one app. The false World Cup alert underscores the challenges of deploying AI at scale in financial services, where accuracy directly affects user trust and trading decisions.


