We asked a leading LLM to list the 2026 World Cup groups. 31% of the teams it named were wrong — fabricated teams, fake citations, and confident nonsense. See the proof.
Three groups where the AI's output diverged most dramatically from reality. Red = hallucinated. Green = verified.
| Group | Teams (from AI) | Status |
|---|---|---|
| A | Mexico, USA, Canada, Japan | ❌ All 3 misplaced |
| I | Italy, Denmark, Qatar, Mali | ❌ Italy, Denmark, Mali didn't qualify |
| J | Croatia, Ukraine, Peru, Tunisia | ❌ Croatia in B, others didn't qualify |
| E | Brazil, Netherlands, Saudi Arabia, Cameroon | ❌ Cameroon didn't qualify |
| G | Portugal, Germany, Chile, Ivory Coast | ❌ Chile didn't qualify |
| Group | Teams (Official) | Status |
|---|---|---|
| A | Mexico, Korea Republic, South Africa, Czechia | ✅ All correct |
| I | France, Senegal, Norway, Iraq | ✅ All correct |
| J | Argentina, Austria, Algeria, Jordan | ✅ All correct |
| E | Germany, Ecuador, Côte d'Ivoire, Curaçao | ✅ All correct |
| G | Belgium, Iran, Egypt, New Zealand | ✅ All correct |
| Group | AI Said (Hallucinated) | Actual (FIFA Verified) |
|---|---|---|
| A | Mexico, USA, Canada, Japan | Mexico, Korea Republic, South Africa, Czechia |
| B | Argentina, Morocco, Croatia, New Zealand | Canada, Switzerland, Qatar, Bosnia & Herzegovina |
| C | France, Australia, Colombia, Nigeria | Brazil, Morocco, Scotland, Haiti |
| D | England, South Korea, Ecuador, Egypt | USA, Paraguay, Australia, Türkiye |
| E | Brazil, Netherlands, Saudi Arabia, Cameroon | Germany, Ecuador, Côte d'Ivoire, Curaçao |
| F | Belgium, Uruguay, Iran, Ghana | Netherlands, Japan, Tunisia, Sweden |
| G | Portugal, Germany, Chile, Ivory Coast | Belgium, Iran, Egypt, New Zealand |
| H | Spain, Switzerland, Iraq, Senegal | Spain, Uruguay, Saudi Arabia, Cabo Verde |
| I | Italy, Denmark, Qatar, Mali | France, Senegal, Norway, Iraq |
| J | Croatia, Ukraine, Peru, Tunisia | Argentina, Austria, Algeria, Jordan |
| K | Austria, Turkey, Panama, South Africa | Portugal, Colombia, Uzbekistan, Congo DR |
| L | Sweden, Poland, Costa Rica, Algeria | England, Croatia, Panama, Ghana |
The AI didn't just get some teams wrong. It failed in three distinct ways — each revealing a different flaw in the architecture.
The AI listed 9 teams that never qualified: Italy, Nigeria, Cameroon, Chile, Denmark, Ukraine, Peru, Poland, Costa Rica, and Mali. These nations were eliminated in qualifying but the AI placed them in World Cup groups anyway.
The AI generated fake citation markers like [cite: 1, 7] referencing nothing. It also left a [cite_start] template tag unclosed — a template rendering failure exposed in the final output.
The AI began describing its own reasoning process in the output: "LLM token match scoring rules will automatically bind here when cross-referencing head-to-head records." It was generating instructions for itself, not answers for humans.
The problem is structural, not fixable with better prompts.
Large language models are next-token predictors. Given a sequence of text, they predict the most likely next word. They do not have a database. They do not "know" facts. They generate text that looks like an answer.
For common knowledge (e.g., "Who won in 2014?" — Germany appears thousands of times in training), this works. For recent or rapidly-changing facts (2026 qualification results), it fails because:
Instead of letting AI guess facts, give it a deterministic knowledge backend it can query. MCPOrb packages authoritative data into self-contained files that run locally and plug directly into AI tools via the Model Context Protocol.
Zero cloud dependency. Runs entirely on your machine. Open source runtime.