Bridging the AI Investment Divide

Illustration via iStock/Dmitrii_Guzhanin

While AI global venture capital investments exceeded $290 billion between 2019 and 2024, less than 1 percent of this funding went to initiatives aimed at social impact. This stark disparity reveals a profound misalignment between AI’s transformative impact potential and its current applications.

This disparity is also reflected in the initiatives available to social innovators looking to deploy AI for impact, which are limited not only in number but in type. A 2024 report by the Schwab Foundation, in collaboration with Microsoft and EY, found that only 7 percent of publicly available and self-reported impact-focused AI initiatives were centered around AI education or skills development. As a result, adoption rates among social innovators and impact enterprises generally lag behind the 78 percent of global companies either using or tangibly exploring the use of AI.

The evidence is clear—AI delivers value for the social sector. We have found, through our work and interactions within the Schwab Foundation and MIT Solve networks, that social innovators are not only deploying AI for social and environmental impact but more importantly, are finding innovative ways to mitigate and offset the technology’s associated risks. However, wider concerns about the AI skills gap, data bias, costs of entry and infrastructure access remain, leaving the question: How can social innovators ethically access and adopt AI capabilities at scale and at a rate that ensures these changemakers do not fall behind?

Social innovators, especially those located in lower- and middle-income regions, need a clear roadmap to increase their awareness and guide their deployment of the technology. As the world faces mounting global challenges, from educational inequity to health care access, redirecting AI investment toward social innovation and other impact applications isn’t just a box ticking exercise; it’s a global strategic imperative.

Current AI Investment Trends Widen Inequality

While it has primarily been touted as a new driver of productivity and financial growth, AI is so much more than a profit-making tool. The current concentration of AI funding in profit-driven applications creates a troubling pattern of widening inequalities. This manifests in three critical ways:

1. Resource Allocation: The focus on commercial applications diverts technical talent and research capacity away from social impact initiatives.
2. Geographic Disparity: Even when it is directed for social impact, funding goes predominantly to those social initiatives originating from high-income countries, creating an innovation gap in low- and middle-income countries where solutions are most needed.
3. Data Bias: This is all happening in an emerging tech landscape that’s already underrepresenting those populations. Leading AI models are largely sourced on datasets from the United States, China, the European Union, and the United Kingdom, perpetuating algorithmic biases against underrepresented populations

AI’s Transformative Potential for Social Good

In our work at the Schwab Foundation and MIT Solve, and through our member networks, we support thousands of social entrepreneurs around the world demonstrating how AI can be deployed to improve people’s lives and drive positive societal transformation—from improving health and learning outcomes, to mitigating the effects of climate change. In addition to the breadth of application, AI helps with scaling impact: solutions in Solve’s portfolio that are powered by AI reach twice as many lives as those not making use of the technology. Similarly, social entrepreneurs in the Schwab Foundation community have reported up to 30 percent internal efficiency gains achieved through their deployment of AI.

Take for example Nigeria’s LifeBank—which uses sophisticated AI algorithms to deliver life-saving blood, oxygen, and equipment to hospitals. Despite having limited access to capital, LifeBank turned to a tech-driven approach, partnering with Google Maps to identify the quickest routes for blood delivery and compiling data from blood banks that can be accessed 24/7. Today, Lifebank serves 3,000 hospitals that meet the health needs of a population of over 40 million people in 15 cities and three countries. For 2024 Schwab Foundation awardee and the founder of LifeBank, Temie Giwa-Tubosun, understanding and leveraging AI became a business imperative to enable the organization to achieve its “large, audacious” impact ambitions. The technology has emerged as a tool that allows them to leapfrog a tough operating landscape and challenges that include infrastructure deficit, technology underinvestment, and access to capital.

Another social enterprise, SXD,
transforms some of the 92 million tons of fabric material wasted each year into valuable new products and eliminates future waste through automated zero-waste AI-generated product design. This approach results in an approximately 80 percent reduction in CO2 emissions and up to 69 percent in material savings relative to conventional methods.

As for Amini’s AI-powered data platform, it provides hyper-accurate data localized to smallholder farmers in areas of Africa where information and internet access are both scarce. Utilizing satellite imagery, and machine learning, Amini aggregates and translates complex raw data into useful information accessible via an API, text message, or WhatsApp for those with no smartphones, providing them with high-quality insights, thereby improving their access to financial resources like agricultural insurance and strengthening their climate resilience.

These stories share a common thread: social entrepreneurs on the frontlines of inequality and climate crises, leveraging AI to create practical, culturally appropriate, and scalable solutions that meet the needs of the communities they are located in. They can develop innovations that are both technically sound and socially acceptable, precisely because they know the problem and local context firsthand.

Deploying AI With Intent: The PRISM Framework

Yet, the technology can remain largely accessible for social innovators despite a strong desire to adopt and deploy its capabilities. Issues with adopting AI can range from a seeming unknowability fueled by an illusion of technical complexity to a lack of internal resources and readiness to a lack of access.

In 2024 the Schwab Foundation, in collaboration with Microsoft and EY, released the PRISM Framework, drawing on insights and lived experiences from social innovator pioneers who have already begun incorporating AI into their work. The framework presents a modular approach, accompanied by real world examples to guide social innovators along their AI adoption journey, regardless of their maturity level. It consists of three layers which along with several interconnected dimensions and metrics through which each social enterprise can evaluate its AI work and proffers actionable recommendations and next steps to take.

1. Layer 1: Impact Mission and Strategy – The foundation of the framework centers the impact mission and strategy of social enterprises. By anchoring AI adoption in their core mission, social innovators can clearly define the “why” behind their pursuit of AI.For instance, SAS Brasil, a 2023 Schwab Foundation awardee and social enterprise working to reduce cervical cancer rates among indigenous women in Brazil, decided that prioritizing ethical adoption is central to their impact mission, especially given their work in underserved regions of the country. This commitment shaped their approach to the deployment of the technology; thus far, they have designed and institutionalized an ethical framework that guides their approach to AI and ensured that the LLMs on which their AI-powered cervical cancer diagnostic tool will be based consists of datasets that are representative of the communities that they serve.
2. Layer 2: Adoption Pathway – This dimension addresses the “how” of AI adoption, offering pathways tailored to an enterprise’s model and readiness. Positioned along a maturity continuum, these pathways range from “conscious tinkering,” suitable for organizations in the early stages of AI exploration, to “AI-first organizations,” which have the internal and external resources to fully integrate AI across every level of their operations. An example is found with High Resolves, an AI-first education technology social enterprise with presence in Asia, Africa, and Latin America, which not only uses AI to deliver personalized learning modules to children but has expanded to include implementing AI tools for other organizations, including those in the social sector.
3. Layer 3: Capabilities and Risk – This layer encourages social enterprises to assess their internal capabilities and potential risks associated with AI adoption by articulating “what” they need to critically consider before deployment. Critical questions that may come up in analysis along this area include: Is our data organized, credible, and free of bias? How complex is the capability we need to deploy? Is there executive and community buy-in? Social enterprises can ensure responsible and effective AI implementation by proactively integrating risk management and capacity-building strategies.

The PRISM Framework also highlights the cost implications associated with deploying the technology – an essential consideration, as high entry costs are often cited by social innovators within our networks as a barrier to AI adoption.

Enabling AI-Powered Social Innovation

While social innovators do the work internally to access and deploy the technology, one thing remains clear: AI’s transformative power must be redirected toward social impact through a collaborative, coordinated, and comprehensive approach. This requires addressing three interconnected systems:

Capital Flows: Venture capital and its yardsticks dominate AI funding, with return on investment measured purely in financial terms, alongside aggressive scale and short-term exit expectations. To incentivize social impact AI investments, we need to rethink how those investment decisions are made and measured. Impact investors and venture philanthropists are well placed to fill this gap, with blended financial models that incorporate social return metrics alongside financial ones and have more extended investment horizons. One of Solve’s investments, Globhe, is a poster child of what profit and purpose can produce. Having raised a seed round, it is generating profits, and the data captured by its fleet of drones is being used to ensure reforestation, roll out renewable energy, ensure disaster preparedness, and more covering areas with 60 million inhabitants.

Knowledge Exchange: There is a world in which tech giants who are funding most AI development can partner with social innovators, the former providing technical power and infrastructure support, while the latter positively influences the development of the technology, ensuring it is equitable, free of bias, and fit for purpose. We see this firsthand when we broker partnerships between innovators and corporations. Take Livox, a software for tablets that enables non-verbal people to communicate and learn. Livox integrated generative AI into their solution to make it easier for people with disabilities to create content. In May 2024, the Brazil-based team was chosen for Solve’s HP AI in Social Impact Award. Their founder said, “By integrating HP’s support into our platform, we’ve been able to reach children with autism who recently faced devastating floods in southern Brazil—an outcome of ongoing climate change.”

LifeBank’s founder adds, “We’ve been able to deploy partnerships over time with major institutions that sort of like do this skill transfer with us. We partnered with IBM for our blockchain products to leverage their technical expertise. This expertise is then transferred to a key member of our organization, who works closely with our team to develop their capacity to manage and build out our blockchain infrastructure and we are on the lookout for more of these kinds of opportunities.”

Creating mechanisms for bilateral knowledge transfer between tech companies and social entrepreneurs can ensure that innovations are both technically sophisticated and contextually appropriate and demonstrate the next generation of tech business models that marry profit with purpose.

Ecosystem Development: Due to systemic inequalities, social entrepreneurs with innovative solutions to critical social challenges in low- and middle-income countries seldom have access to international funding or powerful networks. In fact, research suggests that only 10 percent of grants funding AI-driven work on the United Nations Sustainable Development Goals in recent years went to organizations headquartered in low and middle-income countries. Without the resources, support, and expertise of these ecosystems, those solutions remain unrealized or confined to small-scale implementations. That is why third parties (accelerators, incubators, international organizations) play a critical intermediary role in helping them obtain the partnership and support of tech companies, investors, and public-sector stakeholders. This means moving beyond traditional accelerator models to create sustained, well-resourced partnerships.

The choice before us is stark: We can allow AI to exacerbate existing inequalities, or we can intentionally reshape its trajectory to create a more equitable world. The examples above show us what’s possible. Now, with proper ecosystem support, capital redirection, and committed partnerships between technology companies and social innovators, we can accelerate the transition to a more sustainable, equitable, and just society for all. The time to act is now. The question is not whether we can afford to redirect AI budgets toward social innovation, but whether we can afford not to.

Read more stories by Hala Hanna & Daniel Nowack.