WEBVTT

1
00:00:00.280 --> 00:00:06.400
Regulators. So I think the key
point was really that it's important for us.

2
00:00:06.480 --> 00:00:09.560
It's important for us to have a
referee, just as you have a

3
00:00:09.560 --> 00:00:12.960
referee in a sports game or all
sports games, and that the games are

4
00:00:12.960 --> 00:00:16.920
better for it to ensure that those
you know, the players, pay the

5
00:00:17.000 --> 00:00:23.480
rules play fairly. I think it
is important for similar reasons to have a

6
00:00:23.519 --> 00:00:27.800
regularator regulator, which you can think
of as a referee, to ensure that

7
00:00:28.079 --> 00:00:31.079
companies take actions that are safe and
in the interests of the general public.

8
00:00:31.120 --> 00:00:34.560
Did you tell them that AI has
a double edged sword? Yes, And

9
00:00:34.600 --> 00:00:38.719
what did you mean by that?
Well, what did the metaphor? Yeah?

10
00:00:38.759 --> 00:00:41.799
I mean I mean well, if
you have a sword with two sides,

11
00:00:42.079 --> 00:00:44.759
when you're in one side can get
you even though you're trying to get

12
00:00:44.840 --> 00:00:49.560
something with the other side. What
do you want them to take away?

13
00:00:49.600 --> 00:00:52.280
I think it more important to make
sure your hilt is also you have a

14
00:00:52.320 --> 00:00:56.920
hilt, not a shop thing.
Yea. How do you think they how

15
00:00:56.960 --> 00:00:59.520
do you think they took the message
that you brought to them? Well,

16
00:00:59.679 --> 00:01:03.159
I think it was. It was. It was very civilized discussion, actually

17
00:01:03.520 --> 00:01:07.000
among some of the smartest people in
the world. So I thought it was

18
00:01:07.840 --> 00:01:14.439
Senator Schumer a great service here,
humanity here along with sport of the rest

19
00:01:14.480 --> 00:01:18.280
of the Senate, and I think
we'll I think something good will come of

20
00:01:18.359 --> 00:01:21.519
this this. I think this meeting
may goun on history as being very important

21
00:01:21.560 --> 00:01:23.640
for the future of civilization. Do
you think some legislation is going to come

22
00:01:23.640 --> 00:01:27.000
out of this? Probably? I'm
not sure what the timeframe of that is,

23
00:01:27.040 --> 00:01:33.879
but you know, the legislative process
is not a swift one. But

24
00:01:34.079 --> 00:01:38.079
I think it's I think this will
the things will go in the right direction

25
00:01:38.159 --> 00:01:44.840
generally. Schumer or Leader Schimer did
ask everyone to raise their hands in the

26
00:01:44.920 --> 00:01:48.480
room to see if they were in
favor of AI regulation, and I believe

27
00:01:48.560 --> 00:01:52.079
almost everyone did. Yeah, so
that's a good sign. Do you think

28
00:01:52.120 --> 00:01:55.640
any of the text CEOs left and
change their minds? Do you think they

29
00:01:55.640 --> 00:01:59.000
were changed by anything they hurt from
the Senate side? Well, I can't

30
00:01:59.000 --> 00:02:02.319
read their mind but judgment by the
fact that they put their hands up when

31
00:02:04.400 --> 00:02:07.479
asked if they if they believed that
some regulation is in order. I think

32
00:02:07.560 --> 00:02:12.639
the general I think it's clear that
there's a strong consensus, a woman of

33
00:02:12.680 --> 00:02:15.039
woman consensus that there should be some
AI regulation, that it would be in

34
00:02:15.080 --> 00:02:19.560
the best interests of the people to
do so. And I think we'll probably

35
00:02:19.560 --> 00:02:23.479
see something happen. I don't know
on what timeframe or exactly how it will

36
00:02:23.479 --> 00:02:25.759
manifest itself. But the questions what
were they going to do? I don't

37
00:02:25.800 --> 00:02:32.199
know. I mean, there's there's
clearly we've created regulatory agencies before and actually

38
00:02:32.240 --> 00:02:37.319
just recently. Just believing me the
point that, you know, while our

39
00:02:37.360 --> 00:02:40.840
regulatory agencies are not perfect, and
I deal with regulators on a very frequent

40
00:02:40.840 --> 00:02:49.159
basis with the automotive, you know, communications to starlink and then FA with

41
00:02:49.159 --> 00:02:54.080
with with rockets. So I've had
transpantent interaction with regulators for you know,

42
00:02:54.199 --> 00:03:00.759
a couple of decades at least,
And while regulators are not perfect, there's

43
00:03:00.759 --> 00:03:04.360
no regulatory agency that I'm aware of
that I would I think we should at

44
00:03:04.400 --> 00:03:08.159
the federal level at least that that
we should delete. Yeah, so I

45
00:03:08.199 --> 00:03:14.719
think I or something like that,
I don't know what exactly, Perhaps perhaps

46
00:03:14.759 --> 00:03:19.599
an apartment of AI. Probably,
let's say, I think the probability of

47
00:03:19.599 --> 00:03:23.520
there being some sort of a regulatory
agency that stands on its own, similar

48
00:03:23.560 --> 00:03:25.680
to the FAO FCC is likely at
some point, you think, sample I

49
00:03:25.680 --> 00:03:31.560
think so. Now, the reason
that I've been such an advocate for AI

50
00:03:31.639 --> 00:03:37.599
safety in advance of sort of anything
terrible happening is that I think the consequences

51
00:03:37.719 --> 00:03:43.520
of AI going wrong are our sphere. So we have to be proactive rather

52
00:03:43.520 --> 00:03:49.199
than reactive in the past. If
he takes a and I'm being somewhat late

53
00:03:49.199 --> 00:03:51.879
for speaking of regulators, I'm a
little late for the FA I'm meeting with

54
00:03:51.879 --> 00:03:55.879
the doone hold you up? I'm
sure, but yeah, if you take

55
00:03:55.919 --> 00:04:00.560
the example of say seat belts,
seat belts were opposed by the auto industry

56
00:04:00.599 --> 00:04:03.240
for a very long time, even
though the data was very clear that they're

57
00:04:03.280 --> 00:04:12.599
safe, that they radically improve deaths
and injuries. So you know, we

58
00:04:12.479 --> 00:04:15.000
we don't want to be in that
situation where we're fighting regulations even though you

59
00:04:15.000 --> 00:04:18.959
know there's a safety thing. We
can't wait for millions of people to die

60
00:04:19.040 --> 00:04:24.279
in auto accidents, as you know, like, and it's important to just

61
00:04:24.319 --> 00:04:29.639
elevate the question here. Question is
is really one of civilizational risk. So

62
00:04:29.879 --> 00:04:32.759
just it's it's not like one group
versus another, one group of humans versus

63
00:04:32.759 --> 00:04:35.439
another. It's like, hey,
this is something that's potentially risky for all

64
00:04:35.639 --> 00:04:42.079
humans everywhere, It's very important to
to understand that is there an equivalent of

65
00:04:43.439 --> 00:04:46.040
really, Congress, it's sufficiently ready
to regulate a no, no, no

66
00:04:46.439 --> 00:04:51.399
no. The sequence of events will
not be jumping in at the deep end

67
00:04:51.199 --> 00:04:57.439
and making rules, but it starts
with and with insight. In fact,

68
00:04:57.519 --> 00:05:00.319
this is actually how all of the
regulatory bodies have been formed. Believe.

69
00:05:00.879 --> 00:05:04.199
As you start with a group form
to create insight to understand the situation,

70
00:05:04.920 --> 00:05:10.079
then you have proposed rulemaking, You'll
get some objections from industry or whatever,

71
00:05:10.759 --> 00:05:15.199
and then ultimately you get sort of
consensus on rolemaking. That rulemaking then becomes

72
00:05:15.279 --> 00:05:19.680
a law regulation. What's resedge to
an ordinary person? Sperment actually, thank

73
00:05:19.680 --> 00:05:25.199
you. Another by Mark Zuckerberg