Can you trust AI to research investments?

Hi Scott,

Love your work! Do you have any advice on using Microsoft AI Copilot for investing? I asked it two related questions about AI-chip maker Nvidia’s biggest drawdown in 2018 and got two different answers:

"What was Nvidia’s biggest percentage drawdown in 2018?" → Copilot said -30.82%.
"What was Nvidia’s biggest percentage drawdown in one year?" → Copilot said -50.26% (also in 2018)

This inconsistency has shaken my faith in using AI for information to help me invest better. Thoughts?

Darren

Muchas Gracias Senor Darren!

You’ve perfectly demonstrated why I don’t take financial advice from AI. (And it’s not just Microsoft, I just asked Lucy the same question, and she came back with “a fall of 57%” …  which is also wrong).

Still, some of the smartest minds are pouring millions into building AI models to outsmart other investors and squeeze out profits. Today, short-term algorithmic trading now dominates the stock market, with bots executing thousands of micro-trades for razor-thin gains—far faster than any human can react.

And that’s partly why I ditched stock picking and stuck with trusted long-term low cost indexing—because beating AI at its own game is a losing battle.

Yet here’s my question to you:

How does knowing what Nvidia did in 2018 help you make a buck in 2025?

It doesn’t.

Scott.