AI chatbots from major tech companies are surprisingly bad at predicting soccer match outcomes. A new study tested systems from Google, OpenAI, Anthropic, and Elon Musk’s xAI on Premier League betting, and they all lost money consistently.
The results show these advanced AI models, which can write code and solve complex problems, somehow can’t figure out basic sports predictions. It’s like having a genius who aces every test but can’t guess a coin flip.
Even Grok Can’t Beat Vegas
Researchers had the AI systems place theoretical bets on Premier League matches using real betting odds. Grok, Elon Musk’s “rebellious” AI chatbot, performed worst of all. ChatGPT, Google’s Bard, and Anthropic’s Claude also struggled, consistently picking losing bets.
The AIs failed despite having access to team statistics, player information, and historical data. They seemed to overthink simple decisions, getting confused by complex analysis when sometimes the obvious choice wins.
This highlights a weird gap in AI capabilities. These same systems can diagnose diseases and write legal documents, but ask them to predict which team wins and they’re stumped.
What This Means
The study reveals that AI still struggles with real-world uncertainty and randomness. Sports betting requires intuition and understanding context that pure data analysis can’t capture. While AI excels at patterns in controlled environments, the chaos of human competition remains challenging.
Don’t expect AI sports betting apps anytime soon. Even the smartest chatbots would probably lose your money faster than you could.
