06/27/2025
Have you ever asked ChatGPT a question and gotten an answer that sounded right but wasn’t? 🤔
According to a 2025 article in The Wall Street Journal, AI chatbots like ChatGPT are trained to find conflicting, incorrect or incomplete information about the subject you’re asking about.
That’s why they sometimes “hallucinate” answers, even when they don’t have the facts. ⚠️
The problem? Most AIs aren’t designed to say, “I don’t know.” Instead, they give answers that seem confident—even if they’re wrong.
💡 So, what can you do to help manage the risk of getting inaccurate information?
Here are some tips from the article:
1️⃣ Be specific. Vague questions often lead to vague (and less accurate) answers.
2️⃣ Provide AI with sources. This can help the process.
3️⃣ Double-check anything important. AI can be a great tool for drafts but final copy may need some scrutiny.
As financial professionals, we hear from clients who are experimenting with AI in all sorts of ways—from writing assistance to travel planning.
While these tools can be incredibly helpful, they still require human judgment to be used wisely—which is why we thought this article was worth sharing.