Reid Hoffman: Not Using Frontier AI Models As A Second Opinion Borders On Malpractice

How do you see AI affecting other aspects of healthcare,

and what promise do you still see on the horizon?

I'll start with a bold statement.

If, as a patient or a doctor, for anything serious,

you're not using one or more frontier models

as a second opinion,

my belief is you're bordering on committing malpractice.

And the reason is because part of what these AI systems,

even though many of them

are not specifically trained for medicine,

have ingested trillion-plus words of information,

have tool discovery for going out to the various databases

and also searching.

And so you should be learning

how do I check what I think with it, right?

It doesn't mean that you sublimate

your critical judgment to it.

You could very well have checked it and go,

No, I think you're wrong. I think it's this.

But as a second opinion,

it is bringing superpowers that no human being has.

And medicine in particular tends to be that

what we do when we train doctors and GPs

for a decade in medical school and residencies

and everything else,

is we are trying to program them into a human database of,

Oh, you are exhibiting these symptoms,

so I ask you some questions to get some more data,

and I make a probabilistic diagnosis

of what might be going on.

That is the AI superpower.

And it feels weird because you're like,

Well, wait a minute, but I'm really good.

I went through tons of certification. I've been doing this.

I'm really good at doing it.

It's like, Great, yes you are,

and that's a good thing.

But if you're not using this as a second opinion,

you're making a mistake both as a doctor and as a patient.