Vision
By empowering local human systems with responsible and scalable AI, we can curate unindexed knowledge and deliver critical social services to the world's most vulnerable at an unprecedented scale.
Written by: Andre Heller, Liam Nicoll, Ahad Ali, Shivani Schmulen,
Problem.
In an age defined by technology and seemingly accessible information, there is still a considerable gap in how information is communicated between people and institutions. This “trust gap” undermines effective social services and contributes to the erosion of public institutions. On the most local level, this gap harms individuals by reducing their ability to gain access to a better understanding of the systems that affect their lives, interrogating the power structures around them, and accessing available services.
One of the most striking examples of this in the 21st century was during the COVID-19 global pandemic. Health institutions like the CDC issued general guidelines; however, many people sought additional information related to their personalized circumstances. When the CDC was ill-equipped to respond to these seemingly niche or outlandish questions, a gap was formed. This gap was quickly filled by misinformation and disinformation that ripped through the sociopolitical ecosystem and dramatically impacted the lives of many. While the COVID-19 pandemic is a high-profile case, it is far from an isolated incident.
The most recent UN Global Risk Report identified misinformation and disinformation as the most existential and under-resourced threat facing humanity today. The threat of misinformation and disinformation is intertwined with the breakdown of trust between communities and local institutions. While there is no scaled way to prevent misinformation and disinformation, we can produce outlets that scale accurate and reliable information, countering negative impacts.
Another key component of misinformation and disinformation, stems from the “last-mile” problem. Currently, the internet holds only 20% of all human knowledge, leaving the remaining 80% as unindexed and inaccessible though popular platforms available to people, creating a significant gap in available information. The last-mile disproproationally affects vulnerable peoples and can lead to serious consequences. This gap of information is what creates space for misinformation and disinformation to flourish, since during times of crisis people are desperate for information. This issue is further exacerbated by the fact that profit incentives for western tech and media companies do not align with the public interest of delivering essential information during critical circumstances.
In line with the UN’s report and in an attempt to rectify the last-mile problem, Signpost is dedicated to a world where there is universal access to reliable and accessible information and believes that the path to empowering millions is through bridging the information gap. Through direct, personalized, and humanized access to reliable and accountable information, vulnerable people are able to make their own decisions and act with agency.
Our History
Since its inception, Signpost has recognized the importance of reliable information in humanitarian crises. The goal of Signport is to use cutting-edge technology to connect crisis-affected people with vital resources such as health or legal services. This philosophy was further legitimized by a Randomized Controlled Trial (RCT) in Greece. The RCT identified that information sessions in the form of 1-on-1 communication increased the uptake for humanitarian services and increased the overall trust in institutions. These findings confirmed that when individuals are able to engage in direct forms of communication and receive personalized response information, they feel more confident in local institutions and are more empowered to utilize social services.
Early stages of Signpost programming were delivered manually via social media websites, chat apps, and local practitioners for maximum outreach and accessibility. These platforms offered local service maps where clients could easily access geolocated maps of vetted service providers. Additionally, the platforms provided a channel for two-way communication where clients could ask direct questions and receive personalized responses from a trained local community liaison within 24 hours. Since 2015, Signpost has reached over 20 million people in 30+ countries.
This feedback loop of information not only provided direct information to clients, but it also developed deep connections between local institutions and community members, building trust and bridging the information gap, ultimately setting the foundation for later SIgnpost initiatives.
The early stages of signpost outreach proved to be incredibly impactful; however, it also demonstrated the current limitations in the aid sector, especially related to available resources and personnel. A prime example of this was seen in Afghanistan. A $25 Signpost social media campaign rapidly generated 48,000 support requests; however, all these requests were to be triaged by only 6 staff members. In the end, the team was unable to efficiently respond to all the inquiries, leaving many with less timely and supportive answers than the moment required. This campaign proved that people, especially those in crisis environments, need and are actively seeking access to vital information at a rate that is currently unmeetable by humanitarian organizations.
Theory of Change
-
By strategically integrating human expertise with automated AI processes in one to one communications, we can increase trust in the humanitarian system and uptake of social services leading to a more reliable, trustworthy, and empathetic information ecosystem.
If: We provide clear options for human escalation during sensitive interactions, and use humans to collect and answer information in high-risk, underindexed areas.
Then: We will be able to scale one to one social work communications 10x using agentic AI, while assuring a safety net for critical situations.
Because: AI alone may fail to handle nuanced or sensitive queries, and human-sourced data in hard-to-reach areas is often more accurate and complete than what is available online. This model ensures that the most critical needs are met with a personal touch while leveraging AI for scale.
-
By leveraging AI and automated processes for data acquisition and analysis, we can build and maintain a comprehensive and up-to-date knowledge base with unprecedented speed and efficiency.
If: We implement automated methods for extracting service information from web scraping, surveys, and human knowledge from natural language, and we systematically ingest information from external sources.
Then: We will close information gaps and significantly expand a knowledge base to reach beyond the edges of the internet with minimal manual effort and expense.
Because: This approach allows us to collect, categorize, and update a vast amount of data from a wide variety of sources, including partner organizations, without requiring significant human intervention for every piece of information. The automation of these processes is what allows the system to scale and adapt to a constantly changing information landscape.ext goes here
-
Core Idea: analyze knowledge gaps, measure retrieval adherence, ingest birds-eye risk and information gap data and simulate potential user queries.
Then: We will be able to identify and address deficiencies in our data and system performance before they lead to negative user experiences.
Because: A data-driven approach allows us to move beyond reactive fixes. By actively searching for what we don't know and where our system falls short, we can improve the quality and completeness of our information, making the service more resilient and effective over time. This creates a sustainable model for improvement that is driven by data insights rather than by anecdotal feedback.
-
By creating and sharing technology that is fundamentally open, free, or subsidized and is tailored to the needs of humanitarians, we can dramatically scale information services through a collaborative network of local NGOs, media organizations, and governments.
If: We create technology that is open, free, or subsidized and specifically designed for humanitarian work and mobilize our existing network of over 10,000 humanitarian service partners.
Then: We can scale information services through local partners at a rate not possible for a single organization or a coordinated intervention.
Because: This approach consolidates open data management and fosters true coordination, allowing different actors to work together on a shared technological platform to provide information services more broadly and efficiently.
Signpost's vision is to leverage responsible and scalable AI to eliminate the "last-mile" information gap. By empowering local partners with AI tools, we can create a self-sustaining ecosystem that delivers timely, critical information to the world's most vulnerable. We believe that onboarding each service or information actor to these useful tools will create a snowball effect, building momentum toward a more resilient and responsive humanitarian system for the future.
