A comprehensive test of Apple’s virtual assistant revealed today that Siri correctly identified only 20 out of 58 Super Bowl winners, while repeatedly crediting the Philadelphia Eagles with 33 championships they never won.
Why it matters: Siri’s poor performance on basic sports trivia fundamentally challenges Apple’s recent AI improvements, highlighting significant gaps between its capabilities and those of competing platforms just months before the company’s planned major AI overhaul.
Testing Results: Paul Kafasis of One Foot Tsunami conducted an extensive evaluation of Siri’s sports knowledge, documenting responses to questions about every Super Bowl winner. The results exposed deep flaws in Apple’s current AI implementation, with Siri providing consistently unreliable information across multiple attempts.
- 34% accuracy rate on Super Bowl winners
- 15 consecutive incorrect answers
- Different wrong answers to same question
Competitive Landscape: The test revealed a striking capability gap between Siri and other AI platforms. When Siri struggled with basic queries, alternative platforms demonstrated consistent accuracy:
- ChatGPT provided correct answers
- Google accurately identified winners
- DuckDuckGo showed reliable results
Future Development: While Apple plans to address these shortcomings with a new LLM-powered version of Siri, users will have to wait:
- WWDC 2025 announcement planned
- Spring 2026 expected release
- iOS 19 integration scheduled
Looking Forward: As Apple prepares to unveil its next-generation AI assistant at WWDC 2025, these test results suggest the update cannot come soon enough for users relying on Siri for accurate information.
