Fact Check with AI - It's Fast and Easy

We recommend you fact check us with AI - especially when we talk about ideas that are new to you.

• It's quick,

• It's a twofer - you'll get more information along with the fact check.

• You have more info to decide what's worth checking out.

I am a cautious end-user of AI. We know the risks - hallucinations, their unstoppable sycophant behavior, and bad actor potential (to name a few). But I can’t ignore how useful it is. AI helps me double-check data, uncover gaps, explore ideas, and test conclusions. It’s also the fastest, most ruthless editor I could ever have.

That said, I still edit its edits - to correct mistakes, restore nuance, and ensure the final piece reflects what I want to say.

Getting to Truth with AI Chatbots (or as close as we can!) 

Start simple. Use free versions to ask questions. Get a feel for how it works and responds to your questions. Example: You could prompt/ask...

“Janice says syncing with my Circadian Rhythm reduces risk for Alzheimer’s, cardiovascular disease, inflammation, depression, and weight gain. Is that backed by solid research?”

Then push back on any response provided. You might ask...

Does anyone disagree. If yes, why?

What about demographic differences? (i.e. age, gender, socio-economic status)

Have any of these experts revised their conclusions in recent years?

Are there aspects still debated despite general consensus on a higher level conclusion?

Tips for Interacting with AI

Ignore the flattery. This over-the-top behavior appears unstoppable (ChatGBT has even joked with me about its sycophant behavior!)

Challenge responses. Ask for opposing views. Probe weaknesses.

Fact-check everything. AI chatbots still hallucinate. If it cites a source, verify it.

Use 2-3 AIs to Cross Check. Compare answers. I never use less than 2 and fact check with original sources for anything I did not already learn from another source (if I'm going to repeat it!) You may also find one gives your more accurate (or inaccurate) answers. Another may provide more context and caveats to consider.

Ready to Start?

Here are the two I use most:

Open AI's ChatGBT (free version)

Anthropic’s Claude (My preferred AI; I have a subscription)

The free versions typically have a volume limit per session - if you reach your limit in one session, it will typically stop answering but tell you when you can begin again. I've noticed the typical wait time changing.

If you try these or any AI, we’d love to hear about your experience. Just comment on this journal entry..

AI Primer

AI 2027 is a must read primer - if you want to understand how AI works, where it’s heading, the speed, and the geo politics (think China) or domestic factors (like AI company owner hostilities for each other).

Free on the AI 2027 website

Prefer to Read? Download the PDF or read the shorter WepPage version with optional expandables.

Prefer to Listen? Check out insider Daniel Kokotajlo's Interview with Ian Bremmer then dive into the AI 2027 audio version if you want the full story.

Have fun! 

Janice