Coinbase’s AI-generated news feed mistakenly published a match result hours before kickoff this weekend — a high-profile example of an AI “hallucination” that briefly spread across social media.
What happened
- On Sunday Coinbase sent an alert saying Norway had beaten Brazil 3-2 at MetLife Stadium, with Erling Haaland scoring twice. The game had not yet started, and Coinbase’s own prediction-market page showed the match as delayed.
- Screenshots of the alert circulated on X (formerly Twitter), prompting Coinbase CEO Brian Armstrong to say he was looking into the incident.
Company response
- Max Branzburg, Coinbase’s head of consumer and business products, said the company removed the incorrect story and updated its systems to “avoid these types of inaccuracies in the future.” He praised “the power of AI‑enabled 24/7 insights for trading” while acknowledging the need to “tune” the system.
- Branzburg also noted — with a wink — that Norway ultimately did beat Brazil that day, although the final score was different: Norway won 2-1 in East Rutherford, New Jersey, and Haaland did score twice in the second half.
- Coinbase declined additional comment and directed inquiries to Branzburg’s statement on X.
Why it matters
- The glitch arrives as Coinbase broadens beyond crypto trading into AI-driven products and event prediction markets. In January the company added prediction markets through Kalshi, letting users trade contracts tied to sports, elections, economic data and other real-world events.
- Coinbase has also been rolling out AI-powered wallets and autonomous agents that interact with blockchain applications — tools that aim to automate parts of the crypto experience but also raise the stakes for accuracy and trust.
- The episode underscores a practical risk for platforms using generative AI: even plausible, vivid outputs can be false, and false alerts tied to markets or trading signals can create confusion or reputational risk.
Bottom line
Coinbase says it fixed the false alert and updated safeguards. The company’s push into AI-enabled markets and products continues to offer new capabilities — but this incident is a reminder that firms must pair those capabilities with robust checks to prevent AI-generated misinformation from impacting users.
Read more AI-generated news on: undefined/news
