Reaching goals faster with artificial intelligence

This spring, the Joint Action for Sustainable Development will welcome a new member to the team: a web crawler. The crawler is currently being fed information in the hope that it will soon be able to recognise key actors for the 2030 Agenda independently.

The Joint Action for Sustainable Development is all about bringing together people working on achieving the goals of the 2030 Agenda and a sustainable future for Germany. After all, even the best initiatives, organisations or start-ups can benefit from the chance to network, swap ideas or learn from one another along the way. But, networking can be time-consuming, so what is the most efficient way to connect?

The Joint Action for Sustainable Development, a joint federal and Länder initiative coordinated by the German Council for Sustainable Development (RNE), has found one answer to the problem. Since going live in September 2022, its platform of the same name has listed and brought together associations, civil society, companies, trade unions and more. But until recently, each actor had to be manually researched and invited by a member of the team or had to express an interest in taking part themselves. In future, artificial intelligence (AI) will help speed up this process: from spring this year, a web crawler will trawl the internet for sustainable organisations in Germany.

A practical next step, and one underscored by Finn Wölm, data scientist and co-founder of the Global Goals Directory, who worked on creating the software for the Joint Action: “We have found that many cities have a solid sustainable development strategy in place, but that these rarely make a point of including civil society, business, organisations, initiatives or associations as a matter of course.” Yet it is precisely these actors, according to Wölm, who act as a point of contact between sustainability and the population. A great deal of the potential to include actors more systematically is certainly going untapped at local government level – a fact Sophia von Petersdorff-Campen, who coordinates the Joint Action for the RNE’s office, is happy to confirm: “A preliminary study highlighted the number of organisations and initiatives playing an active role in achieving the 2030 Agenda that are yet to join the wider conversation to discuss their sustainable solutions.”

With the 2030 Agenda, the international community has set itself the goal of achieving global sustainability by 2030. The Agenda details 17 global Sustainable Development Goals (SDGs), which are the blueprint to achieve a sustainable future. The goals include affordable and clean energy for all, eradicating poverty and hunger, and gender equality. The search function on the Joint Action platform allows users to filter organisations according to the individual SDGs the institutions hope to achieve. Moving forward, the preliminary work for this, too, will be automated: “The AI finds out which SDGs each organisation actually contributes to. In future, this will allow us to target how we address individuals in a much more efficient way and invite them to join the Joint Action”, explains Petersdorff-Campen.

Data in almost real time

Founded in 2019, Global Goals Directory is a team of SDG experts and developers who have set themselves the goal of supporting local governments, districts and cities in Germany on their journey to sustainable development. Among other skills, the team specialises in big data, AI and machine learning in the context of the SDGs. “We show local governments their SDG ecosystem. But we don’t just look for stakeholders, we also analyse their contribution to sustainable development, the role they play in society. As such, we can also show where strengths and weaknesses lie”, says Wölm. For example, the team could find a local government has plenty of partners working on SDG 7, renewable energy, but scarcely any on SDG 5, gender equality. The AI generates this data in almost real time, which can then be used to target gaps effectively.

Unlike AI-driven mapping, manual options are time-consuming and costly. “Many local governments just don’t have the resources to map these issues manually”, says Finn Wölm. But it is vital to keep track of the field if we are to achieve sustainable development, something which simply cannot be done by people alone. A project with a similar aim to the Joint Action has already taken place in Canada: the Movement Map. Ten thousand or so organisations were mapped manually over a period of two years to ascertain their contribution to sustainable development. “Except once you’re done, you may as well start all over again”, explains Wölm. The field is just too dynamic for the speed at which humans work.

The AI has an easier time of it: it scans websites for content to classify organisations according to the SDGs, a skill it has been taught over countless examples, having been fed pre-classified texts categorised by SDG. “The AI is able to recognise a pattern from this – where certain words appear together, it acknowledges this as relevant and can apply the pattern to new texts”, says Finn Wölm.

Of course the Joint Action does not go by the information set out on an organisation’s website alone, explains Sophia von Petersdorff-Campen. “We quality-check anyone who registers with the Joint Action and have set criteria in place to make sure an actor upholds basic democratic values, for example, and that they play a recognisable role in improving sustainable development.” In future, the AI should be able to filter this independently too, and the future looks bright for this partnership. Petersdorff-Campen also wants the Joint Action to further promote the conversation on its work, with registered organisations able to actively shape and update their content on the platform.

Artificial intelligence and its potential for sustainable development

Finn Wölm believes AI could also help achieve sustainable development in other areas in the future: “Making political decisions on the basis of data that is three, four or even sometimes as much as ten years old limits you – everything is shown through the lens of the rear-view mirror. AI opens up interesting options for measuring data in real time, for example using satellite imagery.” If you know where lights are on at night, it is easy to measure population size, for example. Or you can collect data on de- and reforestation, biodiversity and even how much CO2 forests store. “Aerial photography makes identifying tree species and their size easy, and you can calculate on the basis of that”, says Wölm.

AI has proven itself even more useful in achieving the SDGs, for example in research, where it can classify research articles and papers to analyse which topics are already being researched and where there are potential gaps in the literature. Another application is legislation, where AI can scan which laws are drafted and passed and how these laws could potentially play a role in achieving the SDGs. After all, a comprehensive overview is a must to hit your goals.