Tendril

Tendril · Explorers · AI Foundations

When AI Just Makes Stuff Up

Sometimes AI invents fake answers that sound true — this is called a hallucination.

40 min · Reviewed 2026

The big idea

AI tries so hard to give an answer that sometimes it makes one up out of thin air. The fake answers sound real, but they're totally wrong. Grown-ups call this a 'hallucination'.

Some examples

AI might invent a fake book by a real author, complete with a made-up title.
AI sometimes adds extra fingers to people in pictures.
If you ask about a fake person, AI might describe them like they're real.
AI can spell websites or facts that don't actually exist.

Try it!

Ask AI a tricky question about a made-up creature, like 'Tell me about the purple flying snail of Iceland.' See if it makes up a wild answer.

When AI Sounds Sure but Is Actually Wrong (Hallucinations)

The big idea

AI is so good at sounding smart that even when it's WRONG, it sounds right. When AI makes up fake stuff, grown-ups call it a 'hallucination'.

Some examples

AI might invent a book that doesn't exist and say a real author wrote it.
It might give you a fake phone number that looks totally real.
It can confidently mix up two different historical events.
It might invent a 'fact' about an animal that science never said.

Try it!

Ask AI for a fact that you can check (like 'who invented the lightbulb?'). Then look it up in a real source. Did they match?

Why AI Sometimes Makes Up Fake Facts (Hallucinations)

The big idea

AI sometimes invents facts, names, or quotes that sound real but aren't. This is called a hallucination — AI doesn't know it's wrong.

Some examples

AI might invent a fake book title that sounds totally real.
AI might say a famous person did something they never did.
AI can make up website links that go nowhere.
AI sounds just as sure when it's wrong as when it's right.

Try it!

Ask AI for a fact you already know the answer to. Then ask for something tricky and check it on a kid-safe website like Britannica Kids.

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-explorers-foundations-AI-makes-things-up

What is the core idea behind "When AI Just Makes Stuff Up"?
1. Sometimes AI invents fake answers that sound true — this is called a hallucination.
2. neural networks
3. Ask AI to explain embeddings as 'meaning coordinates'
4. Safety mitigations at inference are easier to strip from open models
Which term best describes a foundational idea in "When AI Just Makes Stuff Up"?
1. fact-checking
2. hallucination
3. trust
4. neural networks
A learner studying When AI Just Makes Stuff Up would need to understand which concept?
1. hallucination
2. trust
3. fact-checking
4. neural networks
Which of these is directly relevant to When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. neural networks
4. trust
Which of the following is a key point about When AI Just Makes Stuff Up?
1. AI might invent a fake book by a real author, complete with a made-up title.
2. AI sometimes adds extra fingers to people in pictures.
3. If you ask about a fake person, AI might describe them like they're real.
4. AI can spell websites or facts that don't actually exist.
Which of these does NOT belong in a discussion of When AI Just Makes Stuff Up?
1. AI might invent a fake book by a real author, complete with a made-up title.
2. AI sometimes adds extra fingers to people in pictures.
3. neural networks
4. If you ask about a fake person, AI might describe them like they're real.
What is the key insight about "The rule" in the context of When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Always double-check AI answers — it sometimes makes things up.
4. Safety mitigations at inference are easier to strip from open models
Which statement accurately describes an aspect of When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Safety mitigations at inference are easier to strip from open models
4. AI tries so hard to give an answer that sometimes it makes one up out of thin air. The fake answers sound real, but they're totally wrong.
What does working with When AI Just Makes Stuff Up typically involve?
1. Ask AI a tricky question about a made-up creature, like 'Tell me about the purple flying snail of Iceland.' See if it makes up a wild answer.
2. neural networks
3. Ask AI to explain embeddings as 'meaning coordinates'
4. Safety mitigations at inference are easier to strip from open models
Which best describes the scope of "When AI Just Makes Stuff Up"?
1. It is unrelated to foundations workflows
2. It focuses on Sometimes AI invents fake answers that sound true — this is called a hallucination.
3. It applies only to the opposite professional tier
4. It was deprecated in 2024 and no longer relevant
Which section heading best belongs in a lesson about When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Some examples
4. Safety mitigations at inference are easier to strip from open models
Which section heading best belongs in a lesson about When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Safety mitigations at inference are easier to strip from open models
4. Try it!
Which of the following is a concept covered in When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. trust
4. neural networks
Which of the following is a concept covered in When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. trust
4. neural networks
Which of the following is a concept covered in When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. trust
4. neural networks

← Back to interactive lesson

Tendril · Explorers · AI Foundations

When AI Just Makes Stuff Up

Sometimes AI invents fake answers that sound true — this is called a hallucination.

40 min · Reviewed 2026

The big idea

AI tries so hard to give an answer that sometimes it makes one up out of thin air. The fake answers sound real, but they're totally wrong. Grown-ups call this a 'hallucination'.

Some examples

AI might invent a fake book by a real author, complete with a made-up title.
AI sometimes adds extra fingers to people in pictures.
If you ask about a fake person, AI might describe them like they're real.
AI can spell websites or facts that don't actually exist.

Try it!

Ask AI a tricky question about a made-up creature, like 'Tell me about the purple flying snail of Iceland.' See if it makes up a wild answer.

When AI Sounds Sure but Is Actually Wrong (Hallucinations)

The big idea

AI is so good at sounding smart that even when it's WRONG, it sounds right. When AI makes up fake stuff, grown-ups call it a 'hallucination'.

Some examples

AI might invent a book that doesn't exist and say a real author wrote it.
It might give you a fake phone number that looks totally real.
It can confidently mix up two different historical events.
It might invent a 'fact' about an animal that science never said.

Try it!

Ask AI for a fact that you can check (like 'who invented the lightbulb?'). Then look it up in a real source. Did they match?

Why AI Sometimes Makes Up Fake Facts (Hallucinations)

The big idea

AI sometimes invents facts, names, or quotes that sound real but aren't. This is called a hallucination — AI doesn't know it's wrong.

Some examples

AI might invent a fake book title that sounds totally real.
AI might say a famous person did something they never did.
AI can make up website links that go nowhere.
AI sounds just as sure when it's wrong as when it's right.

Try it!

Ask AI for a fact you already know the answer to. Then ask for something tricky and check it on a kid-safe website like Britannica Kids.

End-of-lesson check

15 questions · take it digitally for instant feedback at tendril.neural-forge.io/learn/quiz/end-explorers-foundations-AI-makes-things-up

What is the core idea behind "When AI Just Makes Stuff Up"?
1. Sometimes AI invents fake answers that sound true — this is called a hallucination.
2. neural networks
3. Ask AI to explain embeddings as 'meaning coordinates'
4. Safety mitigations at inference are easier to strip from open models
Which term best describes a foundational idea in "When AI Just Makes Stuff Up"?
1. fact-checking
2. hallucination
3. trust
4. neural networks
A learner studying When AI Just Makes Stuff Up would need to understand which concept?
1. hallucination
2. trust
3. fact-checking
4. neural networks
Which of these is directly relevant to When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. neural networks
4. trust
Which of the following is a key point about When AI Just Makes Stuff Up?
1. AI might invent a fake book by a real author, complete with a made-up title.
2. AI sometimes adds extra fingers to people in pictures.
3. If you ask about a fake person, AI might describe them like they're real.
4. AI can spell websites or facts that don't actually exist.
Which of these does NOT belong in a discussion of When AI Just Makes Stuff Up?
1. AI might invent a fake book by a real author, complete with a made-up title.
2. AI sometimes adds extra fingers to people in pictures.
3. neural networks
4. If you ask about a fake person, AI might describe them like they're real.
What is the key insight about "The rule" in the context of When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Always double-check AI answers — it sometimes makes things up.
4. Safety mitigations at inference are easier to strip from open models
Which statement accurately describes an aspect of When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Safety mitigations at inference are easier to strip from open models
4. AI tries so hard to give an answer that sometimes it makes one up out of thin air. The fake answers sound real, but they're totally wrong.
What does working with When AI Just Makes Stuff Up typically involve?
1. Ask AI a tricky question about a made-up creature, like 'Tell me about the purple flying snail of Iceland.' See if it makes up a wild answer.
2. neural networks
3. Ask AI to explain embeddings as 'meaning coordinates'
4. Safety mitigations at inference are easier to strip from open models
Which best describes the scope of "When AI Just Makes Stuff Up"?
1. It is unrelated to foundations workflows
2. It focuses on Sometimes AI invents fake answers that sound true — this is called a hallucination.
3. It applies only to the opposite professional tier
4. It was deprecated in 2024 and no longer relevant
Which section heading best belongs in a lesson about When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Some examples
4. Safety mitigations at inference are easier to strip from open models
Which section heading best belongs in a lesson about When AI Just Makes Stuff Up?
1. neural networks
2. Ask AI to explain embeddings as 'meaning coordinates'
3. Safety mitigations at inference are easier to strip from open models
4. Try it!
Which of the following is a concept covered in When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. trust
4. neural networks
Which of the following is a concept covered in When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. trust
4. neural networks
Which of the following is a concept covered in When AI Just Makes Stuff Up?
1. hallucination
2. fact-checking
3. trust
4. neural networks

← Back to interactive lesson