In 2025, artificial intelligence revolutionized businesses, transitioning from an emerging tool to improve efficiency to a primary driver of global, cross-sectoral transformation. It holds the unique potential to act as the “great equalizer” and deliver meaningful societal and economic transformation. But to do so, AI must move beyond ad hoc applications and fragmented pilots.
At a high-level roundtable hosted in Davos by global internet and technology company Tencent in partnership with Devex, experts agreed that the next phase of progress depends on deploying AI as a shared tool: one where the same technology driving corporate growth simultaneously builds the economic resilience required for global development. Crucially, this must be done with intentionality to ensure that the deployment of new technology accelerates progress without exacerbating existing divides.
Dowson Tong, senior executive vice president and CEO of the Cloud and Smart Industries Group at Tencent, explained that systemic growth starts by narrowing this gap. “We know this is an X multiplier in terms of productivity if you know how to use it, so it's very critical that we not only bring the technology to the masses but also help them get trained,” so they can leverage the full power of the technology, said Tong.
Moving from this potential to a scaled reality was the central thrust of the conversation. For the 30 experts in the room from the private, public, and social sectors, the goal was to identify how to move past the "Davos stage" and embed social value directly into the corporate engines and technology that will define the next decade of growth. Their discussion articulated a new strategic consensus for inclusive innovation, rooted in the following core insights.
Attendees were unified in their belief that AI can significantly propel progress in areas such as health, education, and agriculture, particularly in under-resourced regions — because it is already happening.
For example, under its drive to create “Tech for Good,” Tencent has moved away from traditional corporate social responsibility toward a model that integrates social value into its core business engine. In practical terms, this means a push to “lower the cost of distributing technology as powerful as AI to empower everybody,” according to Tong. By repurposing its existing technical infrastructure — such as adapting the noise-reduction AI from its conference software to improve hearing aids — the company is focusing on making AI “usable” for those traditionally left behind. “We want to be able to use technology to help solve real-world problems,” added Tong.
Kelly Clements, deputy high commissioner at UNHCR, said AI had been a “game changer” in expanding the refugee agency’s reach. Its Digital Gateway initiative uses AI to provide refugees with 24-hour support, simplifying procedures around registering for health and education services and providing information on rights and available resources. Its use has allowed UNHCR to “deliver faster and better in more locations,” Clements said.
In Africa, Prajna Khanna, global head of sustainability at global internet company Prosus Group and Naspers Limited, shared how the companies received over 1,100 applications to their Tech FoundHER Africa Challenge from women tech entrepreneurs who are using AI to solve for needs in their communities, such as farming assistance or mental health care.
The next stage is to use AI not just to be reactive or to drive efficiencies but to think about how it can transform some of the biggest challenges of our time, said Alain Labrique, director of data, digital health, analytics, and AI at the World Health Organization. Utilizing AI proactively could prevent a wide range of problems before they happen — such as disease outbreaks, bad crop yields, or critical infrastructure fatigue.
To illustrate this idea, Labrique shared how nongenerative AI is being used to detect anomalies in medical scans — from cervical cancer to tuberculosis lesions. This addresses “some of the manpower shortages we have in the global south to be able to accelerate access to care for those people who need it the most,” he explained.
It can also be used at the individual and community levels to provide a better understanding of a person’s own health, promoting behavior change to support the prevention of illnesses such as noncommunicable diseases, Labrique added.
Many at the event recognized the risk of AI excluding those with lower digital literacy levels or limited access to technology. Unless you're very intentional about how the governance is set up, there will be issues related to access, said Clements.
Attendees discussed potential solutions to ensure responsible application and inclusive adoption, advising social impact organizations to carefully consider the use case by answering a series of questions before deployment: Is it the right model? Is the data biased? Is the system that we're looking at appropriate for the context in which it's being deployed?
WHO, Labrique said, is trying to strengthen regulatory capacity to avoid exploitation of data and misinformation via AI. “It’s hard to keep up with the pace with which AI is progressing, but I think that’s really fundamental — to make sure that countries have the ability to, first of all, assess how good and appropriate an AI tool is for the context in which it’s being deployed,” while also ensuring capacity to hold the developers and/or deployers of those systems accountable for public and patient safety, he shared.
It’s fundamental that affected communities maintain a sense of autonomy in the creation of an AI solution: They should be consulted during design and deployment, but also after implementation. This ensures that any AI tool is truly beneficial for the audience it’s intended to assist and meets the needs of the local context.
To ensure equitable access to AI, people must understand how it works — and how it may apply to them. This means refuting the concern that AI is going to replace them in their work and instead explaining how AI can enhance their labor. When AI is positioned as a creative tool in the workplace, people are more confident in using it, said one roundtable participant. Yet, according to research by Accenture, only 11% of organizations are equipped to support colearning between AI and humans. Addressing that disconnect could mean “higher engagement, faster innovation, and improved productivity,” according to the report.
To maximize potential, organizations need an AI strategy, including training backed by research and created in partnership with a multitude of stakeholders. This collaboration should encompass the donors' funding development activity, tech companies building the AI to ensure any solution is fit for purpose, and relevant community voices.
For a meeting taking place in the context of the World Economic Forum, it’s impossible to ignore the raw economic argument for AI. While the private sector traditionally separates commercial growth from social impact, the consensus among attendees was that AI has collapsed that distinction: Solving critical social gaps in health, energy, and infrastructure is no longer a “side project,” but rather a strategic opportunity to expand the application of private products and services.
This shift moves inclusive AI from the periphery of corporate social responsibility to the center of the growth engine. When technology is deployed with local ownership in mind, it creates a “flywheel effect” where local entrepreneurs find efficiencies that provide an innovation boost for local economies. Ground-level ownership is a hotbed for AI innovation itself, attendees noted.
Khanna believes “the next social unicorn” is going to come from companies that can bridge the gap between high-scale AI and traditional forms of social entrepreneurship. The potential for growth is unprecedented if we remain committed to a path of inclusivity — and a vision of AI for all.