AI Early Warning for Disease, Famine at Risk in USAID Tear-Down

Years of efforts to promote a version of AI in Africa and Asia that aligns with the democratic values of the US could end with the Trump administration’s plan to shutter most of USAID, several people familiar with the agency’s operations say.

US Agency for International Development workers had been using AI for multiple programs, including faster detection of tuberculosis through monitoring the sound of coughs, and an early warning system that indicates risks for famine.

Both of those contracts have been canceled, said an agency employee who received a job termination notice, and who spoke on the condition of anonymity. A former USAID contractor also was granted anonymity to speak about the agency’s AI mission.

In another canceled program, USAID was working with Mozilla to promote “responsible computing” in Kenya, South Africa, and other nations, offering future AI specialists in these countries a way to build technologies for “good,” said the agency employee.

The cuts in the Mozilla program put $1.8 million of federal funding at risk, said Steve Azeka, who leads the responsible computing challenge program for the Mozilla Foundation. The program has helped over 30,000 students in the US and abroad, but plans to expand it to Colombia and Ghana could be affected, he said in an email.

“We are currently looking for other alternatives to sustain the program in existing countries (Kenya, India, and South Africa) while prioritizing building a global learning platform,” he said.

USAID was also partnering with organizations from Canada, Germany, and the UK as it built “responsible AI,” along with companies including Amazon Web Services and Nvidia, the former agency employee said. Nvidia declined to comment, and Amazon didn’t respond to a request for comment.

Cutting back or terminating AI programs could enable technology from China and other nations to make inroads into emerging markets and developing countries, said Aubra Anthony, a senior fellow in the technology and international affairs program at the Carnegie Endowment for International Peace, who previously worked for USAID.

“USAID has been working hard to identify ways to support countries as they seek to build their own AI ecosystems, so that innovation can be sort of grown from the ground up,” said Anthony. “It’s just devastating in many ways, and it leaves a vacuum that will, of course, be filled.”

Contracting Confusion

The Trump administration has moved to largely shutter USAID, canceling most of its programs and laying off most of its 10,000 workers. That includes about 1,600 employees who have been fired in the US.

“After a 6 week review we are officially cancelling 83% of the programs at USAID,” Secretary of State Marco Rubio announced March 10 on X.

“The 5200 contracts that are now cancelled spent tens of billions of dollars in ways that did not serve, (and in some cases even harmed), the core national interests of the United States,” Rubio said. “In consultation with Congress, we intend for the remaining 18% of programs we are keeping (approximately 1000) to now be administered more effectively under the State Department.”

Despite Rubio’s statement, Elon Musk’s Department of Government Efficiency (DOGE) only listed about 2,900 USAID canceled contracts as of March 17, adding to the difficulty in confirming exactly which programs have been cut and what’s being kept as the administration moves rapidly to limit the agency’s work. A State Department spokesperson said in an email that critical USAID programs remain active, but didn’t specify if any had AI-related components.

“USAID continues to support the U.S. coordinated, interagency response to the Ebola outbreak in Uganda; to provide lifesaving HIV care and treatment services; to provide emergency assistance in conflict zones; and to support key American strategic partners,” the spokesperson said March 14. “It is the policy of the United States to sustain and enhance America’s global AI dominance in order to promote human flourishing, economic competitiveness, and national security.”

As court battles continue, a federal judge on Tuesday ruled that Musk’s role in the USAID teardown was likely unconstitutional.

Promoting ‘Soft Power’

USAID started in 1961 during the Cold War as a tool of American “soft power,” using its programs in education and healthcare for goodwill among people in different countries. President Donald Trump now says that the agency funnels “massive sums of money to the ridiculous—and, in many cases, malicious—pet projects of entrenched bureaucrats, with next-to-no oversight.”

An inspector general’s report released in the final months of the Biden administration addressing the “top management challenges” facing the agency found USAID wasn’t consistent in following some requirements for cloud computing services. The report also mentioned ongoing weaknesses in the agency’s information systems, risking account misuse and unauthorized access.

USAID rolled out an AI Action Plan in May 2022, six months before the launch of ChatGPT made AI accessible to a wide variety of users. That plan said the agency should commit to promoting responsible AI, invest in digital data infrastructure, and involve partners from developing countries in AI decision-making. The plan has disappeared from the agency’s website.

A lot of USAID’s AI efforts were built around demonstrating a credible and more benign alternative to Chinese technology, said the USAID employee and contractor. Other work included using AI for faster flood and earthquake monitoring in Indonesia, a program whose fate was unclear, and using AI to preserve old Tibetan texts in India, a program that was cut, according to the USAID worker.

“It’s easy to make the argument that the decimation of USAID could ultimately hurt US businesses, it could hurt US actors,” Anthony said. “This just further strengthens the case for countries who I think rightfully are looking to move away from a US-led AI agenda.”

Read More: How Trump’s USAID Dismantling Will Affect Foreign Aid