What does AI’s ability to self-improve mean?

When do we unplug it?

Can we?

We’re racing forward with AI.

But almost nobody is asking what happens if it outruns us.

Enter Eric Schmidt

Eric Schmidt says the real danger isn’t just bad code—it’s bad leadership.

Eric Schmidt is the former CEO of Google. He recently spoke at the PARC Forum at Stanford with a few of his long-time colleagues.

The setting was informal.

The technology was minimal.

The message urgent.

AI as Persuasion Tools

Schmidt warns AI can be used to manipulate people—individually.  

The smarter AI gets. The less aware we will be that this is happening. AI will make us think we are getting what we want so it gets what it wants.

Politicians could create messages crafted to each voter. 

Lies. 

Division. 

All personalized.

2X as deceptive as 2025. Then 4X. Then 20X

Loss of Human Oversight

As AI improves itself, we might not understand how it works. At some point, we can’t stop it—or guide it.

Global Security Risks

Schmidt says rogue actors may use AI for dangerous experiments. Even biological weapons. And there’s no global plan to stop them. That’s the greatest short-term danger.

Long term, we don’t yet know what AI itself is capable of. 

Are we too late?

Schmidt doesn’t think so. 

Now is the time for governments to put in place regulations

Watch Eric Schmidt explain why AI is now a global risk: Watch here