WEBVTT

00:00.680 --> 00:03.880
Artifcial intelligence in service of humankind

00:04.680 --> 00:07.880
At GIZ, we’ve been working with artificial intelligence for many years –

00:08.000 --> 00:12.480
long before large language models such as ChatGPT

00:12.720 --> 00:15.360
attracted worldwide attention.

00:15.480 --> 00:19.560
In our initiative FAIR Forward – Artificial Intelligence for All,

00:19.960 --> 00:22.920
working with partners in countries such as Uganda

00:22.920 --> 00:26.480
and Indonesia, we’ve reached around 600 million people –

00:26.640 --> 00:30.200
for example using AI language assistants in local languages.

00:30.600 --> 00:34.920
After all, people only really benefit from technology

00:35.240 --> 00:37.120
if they can interact with it in their own language.

00:37.240 --> 00:39.680
So over those many years of work

00:39.680 --> 00:41.800
we’ve not only gained a lot of experience

00:41.920 --> 00:43.600
and expertise,

00:43.720 --> 00:46.600
we’ve also established strong networks around the world –

00:46.720 --> 00:49.720
an important basis for our work today.

00:50.480 --> 00:53.960
Of course, it’s also clear that AI presents some real risks –

00:54.200 --> 00:58.440
from data corruption to potential abuse.

00:58.800 --> 01:01.600
That is why it’s so important

01:01.600 --> 01:05.800
to tap into AI’s potential in a targeted and responsible manner.

01:06.160 --> 01:09.760
Used properly, it can help make international

01:09.880 --> 01:12.880
cooperation more effective

01:13.000 --> 01:16.000
and can significantly advance the achievement

01:16.120 --> 01:18.120
of the Sustainable Development Goals.

01:18.240 --> 01:21.080
Our core principle is

01:21.200 --> 01:26.560
that AI must be values-oriented, transparent and inclusive –

01:26.800 --> 01:28.960
while always serving people.

01:29.080 --> 01:31.840
In international cooperation, we support

01:31.840 --> 01:36.320
the countries of the Global South so that they can represent their interests

01:36.440 --> 01:39.640
with self-assurance in the worldwide debate surrounding AI,

01:39.880 --> 01:43.360
and develop their own AI solutions for their own challenges.

01:44.280 --> 01:47.200
Because we are indeed at a turning point.

01:47.320 --> 01:50.240
It’s now that the decisions are being made about who gets to shape

01:50.360 --> 01:53.520
the global AI race – and who gets left behind.

01:54.280 --> 01:56.000
That’s why we’re promoting the foundations

01:56.000 --> 01:58.560
for a fair digital transformation,

01:58.560 --> 02:03.920
which consist of four things: reliable data, clear strategies,

02:04.680 --> 02:09.200
digital infrastructure and, above all, local digital skills.

02:09.840 --> 02:12.320
One example of our work is openIMIS,

02:12.440 --> 02:13.720
an open-source software

02:13.720 --> 02:16.960
for social security that’s already been deployed in many countries.

02:17.440 --> 02:20.720
It helps governments manage benefit payments efficiently

02:20.840 --> 02:25.120
and fairly – from registration to payment.

02:25.680 --> 02:28.160
As an open-source solution, it not only

02:28.160 --> 02:31.240
encourages technological independence and local innovation,

02:31.600 --> 02:34.160
it also promotes transparency,

02:34.280 --> 02:37.280
and it lets people customise and develop things collaboratively.

02:37.880 --> 02:41.440
The result is a digital infrastructure that is effective at local level.

02:42.440 --> 02:45.240
But it's not just about specific AI applications.

02:45.360 --> 02:49.560
We also support the development of sustainable AI strategies –

02:49.760 --> 02:51.440
for example in Kenya and Rwanda.

02:52.120 --> 02:55.400
And there’s one thing always at the heart of what we do:

02:55.640 --> 02:58.640
building up local digital skills.

02:58.760 --> 03:01.920
Because without that, sustainable change is not possible.

03:02.520 --> 03:05.560
It’s essential to have partnerships when applying AI

03:05.680 --> 03:07.960
because AI depends so heavily on data,

03:08.080 --> 03:11.080
context and social conditions.

03:11.520 --> 03:13.840
No player has the knowledge,

03:13.840 --> 03:16.760
data or cultural sensitivity

03:16.880 --> 03:20.800
to develop effective and responsible AI solutions alone.

03:21.680 --> 03:25.880
Solutions only emerge, therefore, through the interaction of many players.

03:26.120 --> 03:27.400
When things go well,

03:27.520 --> 03:31.120
international companies provide expertise and resources,

03:31.240 --> 03:34.840
local start-ups share their understanding of social realities,

03:35.200 --> 03:37.560
governments set the policy framework,

03:37.720 --> 03:41.920
and civil society protects human rights and ethical standards.

03:42.200 --> 03:43.760
That’s the reason why

03:43.760 --> 03:46.400
cooperation with the private sector is so important to us.

03:46.960 --> 03:50.440
It’s private players that hold global technologies and data

03:50.760 --> 03:53.800
on a large scale – and we make all that usable for local needs.

03:54.440 --> 03:57.400
Take this example: together with partners

03:57.520 --> 04:00.400
such as Google, Mozilla, IBM and the Gates Foundation,

04:00.520 --> 04:02.560
we’re making high-quality data sets

04:02.560 --> 04:05.560
available as digital public goods.

04:05.920 --> 04:09.280
Our partner countries can use these to develop their own AI solutions.

04:10.160 --> 04:12.520
In this way, we’re laying the foundations,

04:12.720 --> 04:15.240
not only for the technological strength of AI,

04:15.400 --> 04:17.360
but also for its worldwide accessibility

04:17.560 --> 04:18.920
ensuring that it’s locally effective

04:19.080 --> 04:22.080
and can be co-designed by as many people as possible.

04:22.560 --> 04:24.200
And it’s exactly this that is helping

04:24.200 --> 04:28.160
to make international cooperation even more effective.