WEBVTT
1
00:00:19.860 --> 00:00:22.800
AI creates enormous opportunity for businesses.
2
00:00:22.800 --> 00:00:24.840
It also creates some real challenges.
3
00:00:24.840 --> 00:00:27.930
They need to quickly deploy and scale the technology,
4
00:00:27.930 --> 00:00:28.950
while at the same time,
5
00:00:28.950 --> 00:00:31.620
ensuring it's proficient, safe and equitable,
6
00:00:31.620 --> 00:00:34.800
secure, and compliant with regulation and policy.
7
00:00:34.800 --> 00:00:37.560
GenAI makes it that much more challenging.
8
00:00:37.560 --> 00:00:39.810
If responsible AI is not at the core,
9
00:00:39.810 --> 00:00:42.090
GenAI risks being value-destroying
10
00:00:42.090 --> 00:00:43.930
rather than value-generating.
11
00:00:43.930 --> 00:00:46.500
What are the main challenges companies face
12
00:00:46.500 --> 00:00:48.603
when trying to use AI responsibly?
13
00:00:49.740 --> 00:00:51.450
Well, I think there are three.
14
00:00:51.450 --> 00:00:54.420
First is the pace of technological change.
15
00:00:54.420 --> 00:00:58.530
With AI, the pace of change is so rapid that I think
16
00:00:58.530 --> 00:01:01.410
that the strategies for managing risks
17
00:01:01.410 --> 00:01:05.130
are being outpaced by the level of novelty
18
00:01:05.130 --> 00:01:08.460
and new capabilities that are being built out in AI.
19
00:01:08.460 --> 00:01:12.060
Second is the ambiguous nature of the risks.
20
00:01:12.060 --> 00:01:15.210
It's a set of known risks as well as unknown risks
21
00:01:15.210 --> 00:01:17.850
that I think the traditional strategies
22
00:01:17.850 --> 00:01:21.420
in managing the risks aren't working anymore.
23
00:01:21.420 --> 00:01:25.230
And third, there's new, and an influx of, regulation.
24
00:01:25.230 --> 00:01:27.720
I think there is a lot of questions that remain
25
00:01:27.720 --> 00:01:30.060
as to whether the regulations provide enough guidance
26
00:01:30.060 --> 00:01:32.760
on how to manage the risks properly.
27
00:01:32.760 --> 00:01:34.710
How prepared are companies right now
28
00:01:34.710 --> 00:01:37.080
to manage the risks of AI?
29
00:01:37.080 --> 00:01:39.840
Well, honestly, I think we would all say
30
00:01:39.840 --> 00:01:42.030
that we wish they were even more prepared.
31
00:01:42.030 --> 00:01:46.320
In fact, our research with MIT suggests that
32
00:01:46.320 --> 00:01:50.940
62% of experts that we polled would say that
33
00:01:50.940 --> 00:01:53.280
existing risk management processes
34
00:01:53.280 --> 00:01:55.290
are not expanding fast enough
35
00:01:55.290 --> 00:01:58.440
to incorporate the various risks of AI.
36
00:01:58.440 --> 00:02:02.580
To get this really right is a two- to three-year process.
37
00:02:02.580 --> 00:02:04.380
You have to think through the strategy,
38
00:02:04.380 --> 00:02:07.980
the governance, the processes and tools,
39
00:02:07.980 --> 00:02:10.350
and then work to change your culture
40
00:02:10.350 --> 00:02:11.850
to really be thinking about
41
00:02:11.850 --> 00:02:14.880
how to effectively manage AI risk.
42
00:02:14.880 --> 00:02:16.290
How can companies move
43
00:02:16.290 --> 00:02:19.683
from their responsible AI ambition to execution?
44
00:02:20.880 --> 00:02:25.500
First, create transparency on the risks that you are facing.
45
00:02:25.500 --> 00:02:30.191
Second, you need to invest in risk mitigation tools.
46
00:02:30.191 --> 00:02:32.460
Create transparency on the mitigation tools
47
00:02:32.460 --> 00:02:33.870
that you already have in place,
48
00:02:33.870 --> 00:02:35.970
and make sure that they are uplifted
49
00:02:35.970 --> 00:02:39.300
to be able to capture the AI-induced risks,
50
00:02:39.300 --> 00:02:42.900
like cyber risks, ethical failures, and the like.
51
00:02:42.900 --> 00:02:46.020
And third, stay agile.
52
00:02:46.020 --> 00:02:48.930
As fast as AI and GenAI technology is evolving,
53
00:02:48.930 --> 00:02:51.963
also the risk mitigation around it needs to evolve.