🧠 Why AI can sound right, but be wrong (and how journalists can handle it)

AI doesn’t “know” facts, it predicts likely wording based on patterns. That means it can:

• Get the main idea right
• But insert a wrong detail from a similar context
• Deliver it all with full confidence

Fluency ≠ accuracy.

---

⚠️ Where this bites journalists

Be especially cautious when:

• Explaining technical topics
• Comparing similar concepts
• Summarising complex processes
• Working fast under deadline

These are prime conditions for subtle (and serious) errors.

---

🛠️ How to guard against it

Use these prompt tactics:

1. Force a check

"Verify each factual claim in your answer."

2. Stress-test it

"What part of this is most likely incorrect or misleading?"

3. Break it down

List key claims and explain the evidence for each.

4. Cross-check framing

Are any concepts being confused or mixed up?

---

🧾 Bottom line

Treat AI like a smart but overconfident assistant:

• Useful for speed and structure
• Not reliable without verification

A journalist's job isn’t just to ask better questions; it’s to push for better answers.

🔔Tap the bell icon on this channel to get notified of new updates on this channel (just a few per week).