Kanika Bahl has experience in both the world of frontier AI and the work of addressing global poverty. Until earlier this year, she led Evidence Action, which has delivered health interventions to hundreds of millions of people, and served as a founding member of Anthropic’s Long-Term Benefit Trust.
When Bahl considered where to focus next, she weighed a role in AI safety. Instead, she chose another challenge: building the ecosystem needed to get useful AI tools to people who could benefit from them. She now leads the AI Access Initiative, which aims to reach tens of millions of people in low- and middle-income countries through existing public systems.
Speaking during a live recording of Devex’s Global Progress in the AI Era podcast at Devex Impact House on the sidelines of the 81st session of the United Nations General Assembly, Bahl called that work her “comparative advantage.” Without “active efforts” to spread AI’s benefits, she warned, “the gap between the global south and the developed world will explode.”
The initiative has an early test of its approach in India’s Telangana state, where it has delivered AI-enabled weather forecasts through WhatsApp to 1 million smallholder farmers. Working with the state government ahead of the monsoon season, the team turned technical forecasts into images and voice messages in the Telugu language, then tested them with farmer groups to improve the design.
“Our scale channel is the government,” Bahl said. “We’re not asking farmers to go to a channel they’ve never explored. We’re not asking patients to go outside of their normal system.”
Reaching farmers is one milestone. The larger test is whether the forecasts help them make better decisions and whether the approach can expand through government systems. Bahl said she hopes to reach 10 million farmers within six months, though her team’s more cautious target is 5 million.
Weather forecasting is one of three “big bets” the initiative identified after reviewing 50 ideas for effectiveness, affordability, the potential to reach large populations, and whether artificial intelligence could make a meaningful difference. The others are clinical decision support for health workers and tools that help patients recognize when to seek care.
The AI Access Initiative has secured $20 million in new funding to “supercharge” its work, Bahl said. And it appears that some of the “big bets” the organization has identified are gaining traction among other funders. As Devex reported, the OpenAI Foundation recently announced a $60 million commitment to bring AI-powered weather and crop disease forecasts to 100 million smallholder farmers over three years.
Bahl’s approach draws on years spent working through governments and existing delivery systems at Evidence Action and the Clinton Health Access Initiative. When Devex spoke with her earlier this year, she warned that promising AI projects could end up in “digital graveyards” if their backers failed to address funding, policy, trust, and last-mile delivery.
In healthcare, her team is exploring whether clinical decision support could be incorporated into India’s national telemedicine network. That work remains at an early stage and depends on the government and other partners adopting it.
Bahl also wants the initiative’s work to travel further than its own programs. She said her team plans to make tools and lessons open source, publish playbooks and keep updating its research on promising applications. She also plans to establish an incubator to test “the next set of big bets” that others could take to scale.