AI for skeptics
What does AI really “know?”
- AI knows everything public. It’s read the Web.
- AI knows what’s on your hard drive and in your accounts, if you give it access.
- AI knows things that aren’t true. “Everything and more,” you might say: hallucinations. Including links to sources that don’t exist.
- AI gets things wrong. It misinterprets what it “knows.”
- AI doesn’t know what’s important. If you ask for one fact and the preponderance of evidence is against it, it will surface that fact. It has no judgment. It will confirm your highly suspect conspiracy theory.
Skepticism and truth
The principles for searching out truth have not changed because of AI:
- Be skeptical. Always check original sources — just as you used to do with Google.
- Don’t quote it. Like Wikipedia, it’s a compilation, and can generate a false impression.
- Seek opposing evidence. Just because one source says something is true — and an AI tool quotes it — doesn’t mean that’s the whole truth.
- Be cognizant of bias. AI tools have biases, just like media sources.
- Learn to interpret what you read. Build your science and statistical literacy.
AI plain-English prompts and AI-generated responses are easy to consume.
Don’t swallow them whole.