President Barack Obama on Bureaucracy VS. Moonshots

We're talking about in the specifically,

even in extended intelligence,

you're talking about government, you're talking about

private industry, academia.

Where should the centre of that research live,

if there is a centre?

It's a good question.

I think that MIT would argue it should be in MIT

(laughter)

I think it's, this round is quite shocking, I think,

to the academics because most of the money

and the power of the core computer sciences

and big companies,

and even open AI, that some of our friends made,

that was a non-profit

and not inside an academic institution..

That's sort of surprising right?

Yeah, historically it would have probably been

a group of academics with the government,

so this is a new thing.

But I feel like, as we start to emerge

out of just the computer science mode

to how does this affect society mode,

for instance, just take one of the areas

that we're really interested in is criminal justice

and bail and parole,

and that's probably much better

to have a statistical AI supporting the judge

but then it's not just about whether it's more efficient.

you don't wanna be judged by a machine, right?

So as we start to move out I think it will be interesting

to see who gets involved.

I think as you start to get into the social sciences

and the law and the philosophy,

that becomes more in government than academia

but it is curious,

we, academia, can't compete from a..,

these guys are getting paid millions of dollars

at post-doc levels,

they've got a tremendous amount of resources.

So I think we have to kind of assume..,

but now the military are talking about funding AI

so there may be another player that has a lot of money.

But right now, most of the billion dollar labs,

all of them are really in business.

Well look, I mean we know the guys who are funding them

you know if you talk to Larry, or others,

their general attitude, understandably, is

the last thing we want is a bunch of bureaucrats

that are slowing us down here

as we chase the unicorn out there

Part of the problem that we've seen is that

our general commitment as a society

to basic research has diminished.

Our confidence in collective action has been chipped away.

Partly because of ideology and rhetoric.

So the notion is that if it's government, it's bad

and that's something that I do think needs to be reversed.

Now that requires government

to be more nimble, faster, quicker, smarter.

It's hard in a big democracy with a lot of diverse views,

sometimes to get it moving fast enough

in the direction that something like AI is moving.

It's moving so rapidly

that sometimes government's always playing catch-up.

The analogy that we still use, 50 years later,

when it comes to a great technological achievement

is a moonshot

[Joey] Moonshot

And somebody reminded me, maybe one of you,

that the space program was a half a percent of GDP,

and that doesn't sound like a lot, 0.5% of GDP

but in today's dollars that'd be 80 billion dollars

that we would be spending annually on AI.

Right now we're spending less than a billion.

That undoubtedly will accelerate

but part of what we're going to have to understand is that

if we want the values of a diverse community

represented in these breakthrough technologies

then government funding has to be a part of it.

If government is not part of financing it then

understandably those who pay the piper, call the tune

and all the issues that Joey's raised

about the values embedded in these technologies

end up being potentially lost

or at least not properly debated.

You bring up a really interesting tension there

that Joey you've written about that idea

of innovation happening on the margins or at the edges

and then with the space program

and NASA really centrally governed.

How does that relationship change this kind of development

and thinking about where the transmission

of those ideas can happen?

Well I wanna emphasize that

the way we now think about crowd wisdom essentially

and a bunch of experiments everywhere.

I think that can accelerate rather than impede progress.

As long as everybody's linked together with a sense of

common purpose and responsibility and accountability.

Just to give a very concrete example:

part of our project in precision medicine

is to gather a big enough database of human genomes

from a diverse enough set of Americans.

All kinds of racial types, ethnic types, body types,

you name it, gender.

That instead of financing medical research

where we give the money to Stanford or Harvard

or some other school and they've got their samples

and they're hoarding them and working on it.

You know it's a very linear process.

You now have this entire database

that everybody has access to

and the potential to short-circuit the research process

before you've got promising candidates for treatment

can be hugely accelerated

because people aren't all holding onto their stuff.

That's the power of the internet,

that's the power of connectivity

in the networked world that we live in.

What I've tried to emphasize though

is that just because the government is financing it

and helping to collect the data

it doesn't mean that we hoard it or only the military has it

It's gotta be a top-down approach

but there does have to be some common set of values,

a common architecture, to make sure that the research

is shared by people that..,

it's not monetized by one group rather than another

and there has to be some core principles

that we all agree to.

That's I think an appropriate role

that a group like NIH for example, can play.

I think that if you look at the moonshot

a lot of the value were the tools

that were created in order to do that.

Or if you look at CERN

they've got some esoteric, physics problem

but they invent the web while they're at it.

So I think these mega projects bring together

an interdisciplinary group to solve a problem.

So that's really interesting.

I think one of the problems with standard peer-reviewed,

government funding is it goes out in this hierarchical

pattern that's very politically correct and very rigorous

but it doesn't really get these big ideas going.

And I would say the other thing that you have,

that you're doing well is the open data initiative, right.

So when you're talking about AI you need data,

the government has data.

I helped start a non-profit in Japan after Fukushima

to get citizens to collect data on radiation measurements.

We have 53 million of them, we did well.

So we came to Washington, D.C. and we did a workshop.

We invited the EPA guys and the NSSA guys.

And they had all the data but they hadn't.,

it was open but it wasn't published.

And actually there was data around the White House

that was for national security they didn't publish.

And we invited all those guys

and our guys taught them how to make the kits.

So they walked around and measured the radiation.

Well now that it's public we can release the data

and what's now developed is a bunch of citizen science kids

a lot of them in Japan, now working with the EPA and NSSA

in trying to figure out how you take radiation data,

empower citizens

and we're selling these kits into high schools

and pivoting into air and things like that

and what's interesting is if you get

kids that are sufficiently motivated, kids on the edges,

and give them some sort of interface for this data

and I think that right now you're agencies

are getting much more...closer

[Barack] We're getting clearer

And it used to be that you'd talk to some professor

at some university and give them limited access to the data

but it turns out that kids will figure out

how to use the data

and right now it's mostly visualization

but once we get AI a farmer may be able to go direct

and collect the data and build a model

and use AI to do something

and that's gonna be a lot of how tools get better

but also how the government can interact with those people.