AI Foundations / Guided learning journey
Can you trust an AI-generated claim?
Learn to distinguish documented facts from unsupported claims, test your judgment against fictional evidence, then connect the method to real source-backed research.
First, understand the distinction
An AI assistant can write a convincing statement without evidence. A reliable researcher checks whether the exact statement is supported by the original material.
- Source
- The document or webpage where information originated.
- Evidence
- A specific passage that directly supports or challenges a statement.
- Claim
- A statement we want to assess, separate from its supporting evidence.
- Hypothesis
- A plausible idea that has not been established as fact.
Worked example
If documentation says a fictional security tool supports AWS and Azure, it does not demonstrate that every cloud is supported. You can report the documented environments while marking the broader claim unsupported.
Remember: Unsupported does not necessarily mean disproven. A publication date differs from the date you captured a page. Marketing claims and independent validation are different kinds of evidence.
No model calls, network requests or storage. Results are based on fixed fictional teaching cases, not independently verified real-world claims.