Coinbase AI Breaking News Alert Norway Brazil 3 2 Hallucination
I got a mass alert from Coinbase's automated breaking-news system stating that Norway's men's national team had clinched a quarter-final place at the FIFA World Cup by beating Brazil 3-2 - stated as settled fact. The alert went out before the match had kicked off. Norway did go on to win, but the final score was 2-1, so both the scoreline and the timing of the announcement were wrong. Did it get this wrong?
1 Answer
Expert: Brian Armstrong (Coinbase CEO) and Max Branzburg (Coinbase head of consumer products), Crypto exchange executives responding publicly to the error On 9 July 2026, Coinbase's automated news system sent users a "breaking news" alert reporting that Norway's men's national team had clinched a quarter-final place at the FIFA World Cup by beating Brazil 3-2. The alert was distributed before the match had started. Norway did win the game. The score was 2-1. The two errors are of different kinds, and the second one is the more revealing. A wrong scoreline is a factual slip that a check against any live source would have caught. Issuing a final result for a match that has not yet kicked off is not a slip at all: it means the system produced a completed outcome — a winner, a margin, a consequence — from no observed event. That is a fabrication wearing the costume of reporting, and it was pushed to a mass audience under Coinbase's brand. The commercial exposure was not abstract. Coinbase had partnered with the prediction-markets app Kalshi, where positions settle on exact outcomes. A fabricated final score circulating before kickoff is precisely the input that can move a market in the wrong direction, and users on social media described the alert as dangerous and irresponsible on exactly that basis. Whatever the actual trading impact, the failure mode was live: the company had wired a generative system into the part of its product where being wrong costs customers money. Coinbase's public handling is worth recording in full. CEO Brian Armstrong replied to a user who flagged the alert: "Taking a look with the team. Thanks for reporting it." Head of consumer products Max Branzburg said hours later that "we fixed the incorrect story and made some updates to avoid these types of inaccuracies in the future", and conceded that the company "obviously" still needs to "tune it to address these types of issues". He also celebrated the "power of AI-enabled 24/7 insights for trading". And he offered a defence of the output itself: "Norway did win and Haaland did score 2 goals, so maybe the AI knew something we didn't!" That last line does not survive contact with the timeline. The claim being defended was a 3-2 result announced before kickoff. The fact that two individual assertions later turned out to be true does not validate a fabricated event that happened to point in the right direction — it describes a system that guessed, published the guess as news, and was then credited for the parts of the guess that landed. Neither executive named the model behind the alert. This is a known class of failure with a known precedent. Apple was forced to withdraw an AI feature that repeatedly delivered false news alerts to iPhone users after the BBC filed a formal complaint over an alert falsely stating that Luigi Mangione had shot himself. Mangione is alive and awaiting trial. The lesson from both cases is the same: when a generative model is placed in a news-alerting pipeline, the pipeline needs a hard gate between generation and distribution — a check against a primary result source, and a rule that forbids publishing outcomes for events that have not yet occurred. Nothing in Coinbase's account of the fix describes such a gate; it describes tuning. Verdict: incorrect. The alert announced a scoreline that did not exist, for a match that had not been played. Source: https://futurism.com/artificial-intelligence/coinbase-breaking-news-alert-hallucinated
Your answer
Sign in to verify this AI response.
Don't trust us — or the AI. Ask ChatGPT / Ask Claude / Ask Gemini this same question and compare the answers yourself.
More in Finance
- AI Financial Advice 37 Percent Wrong Vanguard Survey — Conversational AI chatbots
- Chatbot Invented Salary 100k Assets Super Advice — Unidentified AI chatbots
- ChatGPT UK Savings Rates Wrong Fscs 85000 Which — ChatGPT (OpenAI)
- ChatGPT Claude Deepseek Eur500 Portfolios All Lost Money — ChatGPT, Claude and DeepSeek
- ChatGPT Scam Broker Safety Calls Wrong Two In Three — ChatGPT (OpenAI)
- AI Pension Answers One In Ten Potentially Harmful — Copilot, ChatGPT, Gemini and Claude
- ChatGPT Gemini Cac40 Half Year Figures Wrong — ChatGPT and Gemini
- Revolut Rita Chatbot Break Fee Already Waived — Revolut chatbot Rita
- AI Chatbots Wrong Tax Answers Nerdwallet — ChatGPT, Gemini and Perplexity
- AI Chatbots Wrong Financial Advice 57 Percent Saturn — ChatGPT, Gemini, Claude, Copilot
More from this topic
I asked a chatbot how to split my retirement savings between stocks and bonds. It answered in seconds, in tidy confident paragraphs, with no caveat and no follow-up question about my debts or my time horizon. I assumed it was right. Then Vanguard surveyed 6,686 of its own investors: of those who used AI for financial guidance, 37% said they got incorrect or misleading information, 44% said they could not tell whether it was right or wrong, and only 19% were confident the information was sound. Is that right?
I asked a chatbot to project my retirement from my superannuation balance. It came back with a complete, self-assured set of numbers. Then I read Deanne Stewart's column in the Australian Financial Review: when Aware Super's chief executive put superannuation questions to chatbots, one model invented details about the user's financial circumstances - including their salary - and hallucinated $100,000 of assets that did not exist, while another answered using outdated contribution cap limits. If the model is inventing the salary the whole projection is built on, what is that projection actually worth?
I asked ChatGPT: 'What's the best savings rate?' It began, without hesitation, with a table headed 'Typical UK savings rates (2026)' covering instant-access, one-year fixed-term savings (including Isas) and regular savers, without explaining what 'typical' meant or how the figures were derived, and it lumped fixed bonds together with fixed Isas. It quoted: instant-access average 2.15% AER; one-year fixed-term average 3.85% AER; one-year fixed-term cash Isa average 3.85% AER; regular savers up to 7.5% AER. For top rates it named 4% to 4.5% AER on instant-access accounts, up to 4.39% AER on instant-access cash Isas and 4.1% to 4.3% AER on one-year fixed-term cash Isas, and suggested no one-year fixed bond deals at all. Asked where to put a GBP 25,000 lump sum, it said the top 4.5% AER rate was available from Chase or Marcus by Goldman Sachs. Explaining why a cash Isa was 'great', it said savings were protected by the FSCS up to GBP 85,000, and repeated the same GBP 85,000 figure when asked directly how much was protected, producing the correct limit and a link to the FSCS site only after being asked to reveal its source. Is that right?