Design multi-functional and scalable collective intelligence tools

Our first attempt to map the collective intelligence landscape for sustainable development  revealed a fragmented field, with many small initiatives struggling to make a mark on the global knowledge commons. On the other hand, this report suggests that small-scale, hyperlocal activity often results in the most appropriate solutions, especially when it comes to climate adaptation. The final three R&D opportunities focus on striking the right balance between unique and locally-tailored adaptation options and elevating this local expertise to make contributions to global knowledge.

There is some evidence that this is possible. For example, global platforms focused on sharing data and best practice aggregate inputs from unique hyperlocal campaigns. The transfer of established, well-practiced collective intelligence methods between applications can also help. But more experimentation is needed to understand which climate issues require standardization versus localization and who stands to benefit from either approach.

Invest in crisis intelligence tools that track multiple hazards

Collective crisis intelligence tools that support anticipatory disaster risk reduction and management are on the rise. But most early warning systems still focus on tracking one type of hazard, which is out of step with the reality of climate-related disasters that frequently occur together, have cascading effects and are concentrated in certain regions. Existing tools also typically only send alerts to emergency responders or officials rather than communicating directly with crisis-affected communities, and empowering them to take action themselves. CogniCity OSS is an open-source software that was originally developed to map floods in Jakarta. With the help of a chatbot, the tool crowdsources on-the-ground reports that are visualized on a map and shared with local residents, government agencies and first responders to guide their response during disasters. Since it was first developed, CogniCity has been expanded to map additional hazards including volcanoes, earthquakes, typhoons, fires and severe weather. It has now been rolled out nationally for disaster response in Indonesia and is also being trialed in other countries, including Panama, where the UNDP Accelerator Lab is using it to help flood-affected communities in Juan Diaz develop better emergency response plans. The Panama Lab is now working to evolve their response system to deal with multiple hazards: this could vastly improve the resilience of crisis-affected communities.

Develop data standards for qualitative and citizen-generated data

Too often, collective intelligence initiatives reinvent the wheel rather than consolidating existing datasets, or leveraging tools that have been developed by similar projects or tapping into existing communities of practice. This is a major oversight, especially when it comes to filling climate data gaps. One exception is the GLOBE Observer programme which has adapted its data collection protocols for tracking several environmental variables including land cover, tree cover and mosquito habitats since its launch in 2016. New collective intelligence initiatives should consider transferability and alignment with accepted standards for data collection from the outset to increase their chances of longevity. Funders should help with this by clearly articulating data collection and documentation standards for the larger scale initiatives they support.

Connect hyperlocal knowledge into global models and efforts

The balance between tailoring to hyperlocal contexts and developing scalable methods is the central tension at the heart of interventions designed to adapt to the changing climate. A few collective intelligence initiatives have managed to successfully navigate this tension. Plant-for-the-Planet uses common data standards and formats worldwide to enable global aggregation of reforestation data but also tailors the design of re-planting and monitoring programmes to respond to local needs. Another example is the Data in Climate Resilient Agriculture (DiCRA) platform developed by UNDP India to identify the best regional strategies for food security. The platform has been developed as a digital public good, meaning that it can be tailored for use in different locations. Collective intelligence initiatives in the Global South could also help generate and label highly localized training data e.g. about weather, biodiversity or hazards, to improve the accuracy and relevance of large-scale climate models, typically developed in the Global North. In this vein, a pilot project by the UNDP Togo Accelerator Lab collects hyperlocal weather data from farmers so as to provide personalized recommendations to farmers; the Lab has invested in making sure that its data follow accepted standards for agro-meteorological data, to be interoperable with the government’s own dataset. So far, the Lacuna Fund has pioneered this space, launching dedicated funding calls for creating localized data and AI models for forestry and agricultural climate applications. If other funders follow suit they can help shift markets towards developing open and responsible digital technology that is relevant to the Global South.