Overview

The Violence & Impacts Early-Warning System (VIEWS) is an open-source AI-driven platform providing monthly forecasts on the likelihood and severity of armed conflicts worldwide, up to three years ahead. Developed by a research consortium led by PRIO and Uppsala University, VIEWS pools the expertise of renowned researchers across the world to deliver actionable insights, empowering timely interventions to prevent or mitigate conflicts and their adverse humanitarian impacts. All data and forecasting tools are publicly available.

When navigating complex scenarios, the ability to anticipate future developments is essential. Leveraging reliable, credible, and actionable forecasts of emerging violent conflict can enable organizations to effectively mobilize activities and funds to mitigate human suffering and protect human dignity. The Violence & Impacts Early-Warning System (VIEWS) is a state-of-the-art machine learning-based system that forecasts conflicts within and between states globally, focusing on scenarios involving government-affiliated actors and projecting up to three years into the future. The project spearheads cutting-edge research on conflict forecasting and the analysis of its social impacts while actively engaging with key policymakers and stakeholders through dedicated outreach and knowledge transfer initiatives.

VIEWS is jointly hosted by the Peace Research Institute Oslo (PRIO) and Uppsala University while simultaneously engaging affiliated researchers from a dozen additional institutes across the world. The consortium unites experts across disciplines to pool knowledge and develop systems that generate critical insights, facilitating timely and strategic actions to mitigate the impacts of violent conflict.

The system is fully operational and publicly accessible while also undergoing constant development to continually enhance and extend early warning capabilities. In 2024, VIEWS will broaden its scope to forecast non-state violence, where government-affiliated actors are not involved, and one-sided violence against civilians. New tools are also in development to better contextualize and understand the forecasts, enhancing the usability for policymakers, humanitarian organizations and researchers alike. This includes current efforts to estimate uncertainty distributions around all forecasts to enhance the transparency and actionability of early warnings.

VIEWS is deeply committed to FAIR data and open-source principles, ensuring all methodologies, processes, code, forecasts, and performance evaluations are available for external review. These commitments are vital for transparency, credibility, and actionability. In turn, these factors foster trust, encouraging adoption by relevant organizations and advancing the development of peace technologies for the common good.

How does your project support peacebuilding and/or conflict resolution efforts in the context of a humanitarian crisis or developmental context?

VIEWS is uniquely positioned to integrate seamlessly with existing expert-based conflict tracking systems. Our risk assessments provide a comprehensive overview of the future conflict landscape, serving as an ideal foundation for deeper analysis by traditional experts who, in turn, can offer nuanced insights into local conflict dynamics and account for unique factors that data-driven models may overlook.

The forecasts ensure that (1) protracted conflicts slowly fading from news media headlines remain on the radar of policymakers and practitioners; (2) conflicts at low but non-negligible risk of devastating outcomes gain international attention; and (3) obscure conflict patterns exacerbated by compound risks are highlighted. Consequently, our forecasts help combat cognitive bias in traditional expert assessments. This synergy allows for numerous complementary rather than competitive perspectives, enabling multiple touchpoints to trigger both conflict mitigation and peacebuilding actions. The open access to our forecasts can further empower local populations and institutions to promote their own safety and security to spearhead locally owned initiatives supporting durable peace while holding decision-makers accountable to act on issued alerts.

VIEWS is expanding to encompass a broader definition of conflict. We are developing models forecasting multifaceted conflict impacts, such as migration patterns, child mortality, malnutrition and vaccination rates.

In what ways does your project contribute to the existing PeaceTech ecosystem and research efforts in a compelling way?

In 2021, VIEWS pushed the boundaries of conflict early warning capabilities by shifting from predicting the probability of conflict-related fatalities surpassing a threshold to precise point predictions of fatality numbers. We expanded beyond forecasting relative changes in conflict risk or baselines while simultaneously providing more reliable, actionable, and credible predictive outputs through commitments to FAIR data and open-source principles. All methodologies, processes, code, forecasts, and performance evaluations are available for external review, transparently advancing conflict forecasting and developing PeaceTech as common goods.

The point predictions are harnessable for wide ranges of executable decision-making, including risk early warning and anticipatory preparations that meeting thresholds or changes from relative baselines overlook. Finely detailed forecasts enable post-processing to any desired format, unit, and aggregation level. Combined with the 36-month forecast window, VIEWS facilitates long-term strategic planning with extensive lead-times. Automated processes handle massive amounts of data to generate crucial responses for real-world events.

VIEWS is again redefining scopes of inquiry by incorporating conflict forecasting as a distribution of possible outcomes to facilitate a wider range of alerts. Currently captured structural conflict risks will be transformed into forecasts of the adverse impact of armed conflict on humanitarian outcomes, complimenting the operational conflict forecasting.

With the award funds, how would you expand the scope and applicability of your project or research beyond its initial pilot?

Ensuring effective scientific communication and knowledge transfer for research outputs is often hindered by funding that is restricted to specific activities. The continuous development and maintenance of the VIEWS system, for example, is secured until 2027, but securing funding for activities related to training, educational material, and public dissemination remains a challenge. This induces inertia of users, and limits capacities to effectively aid those taking advantage of our resources.

In the spirit of FAIR principles, the award funds would be used to develop technical and non-technical capacity-training material to disseminate VIEWS’ predictive outputs as a common good while ensuring correct, responsible, and ethical use. It would enable us to design a capacity-building catalogue of modules for teams seeking to pilot or integrate our forecasts into their daily operations, including technical implementation directions for internal data solutions and end-user guides. Furthermore, VIEWS would develop a knowledge hub on approaches to conflict forecasting and methodology, expanding upon knowledge from organising prediction competitions. This includes the provision of material concerning the pitfalls and promises of using predictive analytics for anticipatory action – indeed, promises of PeaceTech can only spread as far as they are leverageable.

How does your work leverage collaborations and partnerships to unlock new opportunities and maximize impact?

Our research is conducted in partnership with renowned research institutes across the globe, fostering a dynamic and interdisciplinary exchange of knowledge and expertise that not only enriches our system but also propels the field of conflict forecasting forward. In close collaboration with research institutes such as Karolinska Institutet, the Varieties of Democracy Institute at University of Gothenburg, the Centre for Research on the Epidemiology of Disasters, Universitat Autònoma de Barcelona, International Security and Development Center, and University of Pittsburgh, we ensure that our methodologies are robust, our data is comprehensive, and our models are validated by the brightest minds in the field. This motivation also rings true for our recurring prediction challenges, the latter of which brought together over a dozen research teams from a range of academic, research and governmental institutions to develop better methodologies to predict conflict with uncertainty.

Moreover, VIEWS serves as a unique bridge between research and practice. We actively collaborate with policymakers and practitioners that seek to leverage our data, improve internal forecasting capabilities, or pilot the use of predictive analytics to support anticipatory action and timely responses to humanitarian crises. To this end, we have partnered with a range of IGOs, foreign ministries, and multilateral organisations through the years to develop forecasting products addressing the most pressing needs of each organisation, whilst promoting transparent, responsible, and ethical use of AI. We are also a key partner to the Complex Risk Analytics Fund (CRAF’d), a large UN-led multilateral ecosystem that seeks to leverage interconnectedness to save lives and prevent duplication of technological advancements.