WEBVTT

1
00:00:00.180 --> 00:00:01.620
- The story will be about

2
00:00:01.620 --> 00:00:04.740
how AI powered investing
took trillions of dollars

3
00:00:04.740 --> 00:00:07.710
of assets from passive
funds and from active funds,

4
00:00:07.710 --> 00:00:10.350
because I think that fundamentally
using AI as the future.

5
00:00:19.050 --> 00:00:20.550
- Welcome to AI Factor, coming

6
00:00:20.550 --> 00:00:22.500
to you from the NASDAQ market site.

7
00:00:22.500 --> 00:00:24.420
Today we're talking with Doug Clinton, CEO

8
00:00:24.420 --> 00:00:25.770
of Intelligent Alpha about

9
00:00:25.770 --> 00:00:28.530
how AI is changing
investing. Welcome, Doug.

10
00:00:29.400 --> 00:00:30.990
- Great to be here. Thank you.
- Some 61%

11
00:00:30.990 --> 00:00:32.760
of US adults own stock according

12
00:00:32.760 --> 00:00:34.770
to a recent Gallup poll when it comes

13
00:00:34.770 --> 00:00:36.390
to giving people investment advice

14
00:00:36.390 --> 00:00:38.160
and strategy, how do you see AI

15
00:00:38.160 --> 00:00:39.780
as changing the playing field?

16
00:00:39.780 --> 00:00:41.730
- I think AI will change
the playing field in in

17
00:00:41.730 --> 00:00:43.050
probably numerous ways.

18
00:00:43.050 --> 00:00:44.910
And the fun thing about AI is

19
00:00:44.910 --> 00:00:47.460
that whatever we think
is going to happen today,

20
00:00:47.460 --> 00:00:49.380
sitting here right now, Rossio,

21
00:00:49.380 --> 00:00:51.540
we're probably gonna be drastically wrong.

22
00:00:51.540 --> 00:00:55.380
But I think the big thing
that AI brings to investing

23
00:00:55.380 --> 00:00:57.810
that we don't have right now with humans,

24
00:00:57.810 --> 00:01:02.790
with passive indexes is this
unemotional intelligence

25
00:01:02.790 --> 00:01:04.680
that is AI superpower.

26
00:01:04.680 --> 00:01:07.110
And so at Intelligent Alpha,
we've been experimenting

27
00:01:07.110 --> 00:01:09.750
with using these modern
large language models

28
00:01:09.750 --> 00:01:13.860
for over a year now to
build investment portfolios.

29
00:01:13.860 --> 00:01:17.520
And we've just launched our
first fund on the nasdaq.

30
00:01:17.520 --> 00:01:20.160
And what we've seen in
terms of why AI seems

31
00:01:20.160 --> 00:01:23.760
to perform well, at least in
our estimation from our tests,

32
00:01:23.760 --> 00:01:26.640
is that it doesn't get hung up
when the market's going down.

33
00:01:26.640 --> 00:01:29.070
It doesn't get nervous, or
when the market is booming,

34
00:01:29.070 --> 00:01:31.445
it doesn't feel like it's
missing out and doesn't get fomo.

35
00:01:31.445 --> 00:01:35.280
And so I think the
really key thing about AI

36
00:01:35.280 --> 00:01:37.560
and what it brings into
the market that is new

37
00:01:37.560 --> 00:01:40.740
and unique is just that
unemotional intelligence

38
00:01:40.740 --> 00:01:42.960
- From the investor side,

39
00:01:42.960 --> 00:01:46.680
I think there's some anxieties about AI

40
00:01:46.680 --> 00:01:48.810
and maybe anxieties about some of

41
00:01:48.810 --> 00:01:50.910
that unemotional intelligence
that you're talking about.

42
00:01:50.910 --> 00:01:52.710
What would you say to people to kind

43
00:01:52.710 --> 00:01:54.960
of soothe those concerns?

44
00:01:54.960 --> 00:01:57.060
- There are a lot of different
ways that people can use

45
00:01:57.060 --> 00:01:58.590
and will use ai.

46
00:01:58.590 --> 00:02:00.720
I think to invest in the future.

47
00:02:00.720 --> 00:02:02.010
The way that we do it

48
00:02:02.010 --> 00:02:05.190
and how we kind of address
this concern with investors is

49
00:02:05.190 --> 00:02:07.440
that we still do have a
human oversight layer.

50
00:02:07.440 --> 00:02:10.890
So for us ai, our AI investment
committee we call it,

51
00:02:10.890 --> 00:02:13.950
which is using three different
large language models,

52
00:02:13.950 --> 00:02:15.690
we still have the portfolios

53
00:02:15.690 --> 00:02:20.550
that those ais create
reviewed ultimately by humans.

54
00:02:20.550 --> 00:02:22.950
So a human will say,
okay, does this strategy,

55
00:02:22.950 --> 00:02:24.600
does the portfolio that
the AI put together,

56
00:02:24.600 --> 00:02:26.580
does it make sense in
terms of what we're trying

57
00:02:26.580 --> 00:02:28.590
to create for our investors?

58
00:02:28.590 --> 00:02:31.530
And does it violate potentially
any regulatory issues

59
00:02:31.530 --> 00:02:33.180
for funds that we might manage?

60
00:02:33.180 --> 00:02:35.130
All those things are still seen,

61
00:02:35.130 --> 00:02:36.360
at least at the human level.

62
00:02:36.360 --> 00:02:39.000
And I think in a lot of cases with ai,

63
00:02:39.000 --> 00:02:41.070
we probably still will have some level

64
00:02:41.070 --> 00:02:42.960
of human oversight just to make sure

65
00:02:42.960 --> 00:02:44.880
that there's no mistakes.

66
00:02:44.880 --> 00:02:47.040
The thing where I think
people get really scared

67
00:02:47.040 --> 00:02:49.680
with AI is when they think about AI sort

68
00:02:49.680 --> 00:02:51.210
of being unsupervised

69
00:02:51.210 --> 00:02:54.240
and just free to trade
as as it might, we get

70
00:02:54.240 --> 00:02:56.820
that question a lot with our portfolios

71
00:02:56.820 --> 00:02:58.800
and there are some other investors that

72
00:02:58.800 --> 00:03:01.180
that will probably try
to use that approach

73
00:03:01.180 --> 00:03:03.100
and maybe they'll find ways to do it

74
00:03:03.100 --> 00:03:04.840
with comfort in the future.

75
00:03:04.840 --> 00:03:08.020
But I think really that
human plus AI element, having

76
00:03:08.020 --> 00:03:10.360
that oversight layer will be
how a lot of things get done,

77
00:03:10.360 --> 00:03:12.070
at least for the next decade.

78
00:03:12.070 --> 00:03:13.390
- That makes sense. I think

79
00:03:13.390 --> 00:03:15.430
that's about the stage
that we're at right now.

80
00:03:15.430 --> 00:03:17.080
Do you see in the future ever getting

81
00:03:17.080 --> 00:03:18.340
rid of that human layer?

82
00:03:18.340 --> 00:03:21.970
- I think over time as we
learn to trust the system more

83
00:03:21.970 --> 00:03:25.480
and understand maybe how it
thinks a little bit better,

84
00:03:25.480 --> 00:03:28.390
I think you could see the
human layer probably pull back

85
00:03:28.390 --> 00:03:30.520
some, reduce some.

86
00:03:30.520 --> 00:03:31.750
I don't know that we'd ever want

87
00:03:31.750 --> 00:03:33.610
to necessarily fully eliminate it

88
00:03:33.610 --> 00:03:37.210
because it does seem like
such a useful step, right?

89
00:03:37.210 --> 00:03:38.650
If you're trying to create some sort

90
00:03:38.650 --> 00:03:43.030
of a high frequency strategy
where speed is of the essence,

91
00:03:43.030 --> 00:03:45.370
I think that's a very specific use case

92
00:03:45.370 --> 00:03:47.530
where the human layer would be, you know,

93
00:03:47.530 --> 00:03:48.580
detrimental essentially,

94
00:03:48.580 --> 00:03:50.980
because it would pull
away your speed advantage.

95
00:03:50.980 --> 00:03:53.620
So I do think there will
be cases where things

96
00:03:53.620 --> 00:03:55.570
that orient more towards
speed, you'll want

97
00:03:55.570 --> 00:03:57.190
to eliminate the human layer.

98
00:03:57.190 --> 00:04:00.070
But things that, you know,
as what we do in, in terms

99
00:04:00.070 --> 00:04:02.320
of the portfolios we
build where they're meant

100
00:04:02.320 --> 00:04:05.140
to be longer term oriented
investment portfolios,

101
00:04:05.140 --> 00:04:07.600
I think having that human
layer will probably survive

102
00:04:07.600 --> 00:04:10.060
for a while, even if it does get reduced.

103
00:04:10.060 --> 00:04:11.080
- And so how is the way

104
00:04:11.080 --> 00:04:14.650
that you're deploying AI at
Intelligent Alpha different from

105
00:04:14.650 --> 00:04:16.810
how other strategists

106
00:04:16.810 --> 00:04:19.840
or investment firms are using AI today?

107
00:04:19.840 --> 00:04:23.410
- We think about AI as a
tale of sort of two worlds,

108
00:04:23.410 --> 00:04:27.550
and the world split in fall of 2022,

109
00:04:27.550 --> 00:04:30.370
and that's when Chachi PT was launched.

110
00:04:30.370 --> 00:04:32.950
And I think where you
can kind of categorize

111
00:04:32.950 --> 00:04:36.310
that sep