1
00:00:00,000 --> 00:00:03,917
Good afternoon. I'd like to thank
France for convening this meeting and the

2
00:00:04,000 --> 00:00:06,717
members of the council
for inviting me to speak.

3
00:00:06,800 --> 00:00:11,157
AI is advancing very quickly. Four years
ago, it could barely write a line of

4
00:00:11,240 --> 00:00:15,077
code or solve a high school math problem.
Today, it writes most of the

5
00:00:15,160 --> 00:00:19,317
code at Anthropic and solves
world-famous unsolved math problems. The

6
00:00:19,400 --> 00:00:21,357
trajectory only needs to continue for a

7
00:00:21,440 --> 00:00:24,277
tiny period longer, 1 or 2 years, maybe

8
00:00:24,360 --> 00:00:26,357
less, to reach what I've called a

9
00:00:26,440 --> 00:00:29,037
country of geniuses in the data center.

10
00:00:29,120 --> 00:00:33,757
This rapid pace of progress will produce
incredible benefits for the world. Just

11
00:00:33,840 --> 00:00:37,317
today, Anthropic announced the Claude-led
discovery of a new molecular

12
00:00:37,400 --> 00:00:41,797
machine. It's a preliminary finding,
but it may constitute a new gene-editing

13
00:00:41,880 --> 00:00:46,157
mechanism that could have applications
in in gene therapy. Claude led the

14
00:00:46,240 --> 00:00:48,197
literature review, theorizing, and

15
00:00:48,280 --> 00:00:53,357
experimental design, and humans ran
and verified the experiments in the lab.

16
00:00:53,440 --> 00:00:57,637
We believe this type of application will
accelerate rapidly in the years ahead

17
00:00:57,720 --> 00:01:01,597
and that we can make profound progress
in the near future on diseases that have

18
00:01:01,680 --> 00:01:04,317
plagued humanity for millennia.

19
00:01:04,400 --> 00:01:06,557
But these remarkable capabilities also

20
00:01:06,640 --> 00:01:09,157
come with risks that must be managed.

21
00:01:09,240 --> 00:01:13,917
I see two main categories of risk.
The first is misuse of AI models. For

22
00:01:14,000 --> 00:01:18,557
example, misuse by bioterrorists
to create biological weapons. The second

23
00:01:18,640 --> 00:01:22,597
risk is loss of control, model
capabilities accelerating beyond

24
00:01:22,680 --> 00:01:25,037
developers' ability to control them.

25
00:01:25,120 --> 00:01:29,517
These questions are global in scope and
concern all of humanity. They are grave

26
00:01:29,600 --> 00:01:33,437
enough that I believe they are important
questions for this council, not just for

27
00:01:33,520 --> 00:01:38,357
economic or ICT ministries.
If managed poorly, I even believe that

28
00:01:38,440 --> 00:01:41,477
AI could be a risk to humanity as a whole.

29
00:01:41,560 --> 00:01:45,157
At Anthropic, we believe we have
the best track record in the industry for

30
00:01:45,240 --> 00:01:49,117
preventing these risks,
both operationally and in our research.

31
00:01:49,200 --> 00:01:53,077
We test every model for dangerous
capabilities in areas like cybersecurity

32
00:01:53,160 --> 00:01:58,077
and biology. We put strong safeguards
on our most capable models, and we publish

33
00:01:58,160 --> 00:02:02,477
our work in these areas.
We have pioneered research in areas like

34
00:02:02,560 --> 00:02:06,157
mechanistic interpretability
and alignment science that improve our

35
00:02:06,240 --> 00:02:09,317
ability to steer and control our models.

36
00:02:09,400 --> 00:02:13,837
We also have unparalleled transparency
practices releasing risk reports that

37
00:02:13,920 --> 00:02:15,917
run to hundreds of pages. But we

38
00:02:16,000 --> 00:02:18,477
understand that as AI models become more

39
00:02:18,560 --> 00:02:21,117
powerful, even more stringent standards

40
00:02:21,200 --> 00:02:23,477
of safety will be required and we are

41
00:02:23,560 --> 00:02:26,837
always aiming to improve our safety and

42
00:02:26,920 --> 00:02:31,437
alignment practices, our abilities,
our ability to prevent misuse and we are

43
00:02:31,520 --> 00:02:33,717
always striving to do better. We will

44
00:02:33,800 --> 00:02:36,597
slow down as much as necessary in order

45
00:02:36,680 --> 00:02:41,877
to make sure that every successive AI
technology that we release is actually

46
00:02:41,960 --> 00:02:45,477
safe. But regardless
of what we do, managing

47
00:02:45,560 --> 00:02:49,357
these risks is ultimately bigger
than any one company and it has an

48
00:02:49,440 --> 00:02:52,237
industry-wide and global scale.

49
00:02:52,320 --> 00:02:54,317
On September 12th, I called for a

50
00:02:54,400 --> 00:02:56,197
collaborative effort to pace the

51
00:02:56,280 --> 00:02:59,277
frontier, not to stop AI progress, but

52
00:02:59,360 --> 00:03:01,477
to work together to make sure the

53
00:03:01,560 --> 00:03:03,717
industry and the world can proceed at a

54
00:03:03,800 --> 00:03:08,717
rate where we can be sure that safety
keeps up with capability for everyone in

55
00:03:08,800 --> 00:03:12,917
the industry. This call
had three parts. First, we

56
00:03:13,000 --> 00:03:15,357
committed to embed external evaluators

57
00:03:15,440 --> 00:03:17,397
inside Anthropic with employee-like

58
00:03:17,480 --> 00:03:20,477
access, similar to a food inspector and

59
00:03:20,560 --> 00:03:22,557
and we recommended that other companies

60
00:03:22,640 --> 00:03:28,237
across the world do the same. Some
have already agreed uh uh to adopt this

61
00:03:28,320 --> 00:03:31,877
measure. Second,
we called for cooperation across

62
00:03:31,960 --> 00:03:36,437
the industry to set standards
and modulate the pace of progress. And

63
00:03:36,520 --> 00:03:41,197
finally, we called for global
cooperation across the world between

64
00:03:41,280 --> 00:03:46,077
governments to set international
standards. It is this last area that

65
00:03:46,160 --> 00:03:50,757
most squarely falls within this
council's remit. I want to offer three

66
00:03:50,840 --> 00:03:55,117
ideas in this First,
we should begin with narrow

67
00:03:55,200 --> 00:03:56,877
agreements that every member can

68
00:03:56,960 --> 00:04:00,437
support, such as a ban on using AI to

69
00:04:00,520 --> 00:04:03,197
make biological weapons or permitting

70
00:04:03,280 --> 00:04:06,437
your AI technologies to be used

71
00:04:06,520 --> 00:04:09,317
AI technologies to be used to make

72
00:04:09,400 --> 00:04:12,357
biological weapons. Second, we should

73
00:04:12,440 --> 00:04:17,037
build evaluation and verification
systems that keep pace with AI

74
00:04:17,120 --> 00:04:21,037
development so that states can
have visibility into frontier model

75
00:04:21,120 --> 00:04:23,997
capability and can verify each other's

76
00:04:24,080 --> 00:04:26,797
commitments. Third, we should establish

77
00:04:26,880 --> 00:04:29,357
common global standards for testing AI

78
00:04:29,440 --> 00:04:32,477
models for loss of control risks and

79
00:04:32,560 --> 00:04:35,637
misuse risks and a notification system

80
00:04:35,720 --> 00:04:40,277
for AI incidents that are
significant to global security.

81
00:04:40,360 --> 00:04:44,597
I believe that this is the most
important global security issue facing

82
00:04:44,680 --> 00:04:49,477
the world today. For all the differences
of the people in this room, we must put

83
00:04:49,560 --> 00:04:51,557
aside those differences in order to

84
00:04:51,640 --> 00:04:54,597
confront this global opportunity and

85
00:04:54,680 --> 00:04:59,517
global threat that is being presented
to us at the same time. No leader, no

86
00:04:59,600 --> 00:05:02,597
company, and no nation
can manage this alone.

87
00:05:02,680 --> 00:05:04,477
We commit to working with governments in

88
00:05:04,560 --> 00:05:07,157
this room on this urgent work.

89
00:05:07,240 --> 00:05:10,320
Thank you for your attention.