Funders commit $1B toward developing AI tools for frontline workers

Key Points

• The funders, including the Gates Foundation and Ballmer Group, committed $1 billion over 15 years to develop AI tools for public defenders, parole officers, social workers and other frontline workers.
• They launched NextLadder Ventures, helmed by former Gates Foundation economic mobility director Ryan Rippel, to provide grants and investments to nonprofits and for-profits building these AI solutions.
• The initiative partners with AI firm Anthropic, which will contribute technical expertise, grant access to its Claude model and invest about $1.5 million annually to support grantees.
• The coalition aims to spur economic mobility by matching vulnerable populations with resources and streamlining casework, while emphasizing trustworthy AI and frontline worker feedback.

A coalition of funders, including the Gates Foundation and Ballmer Group, will spend $1 billion over 15 years to help develop artificial intelligence tools for public defenders, parole officers, social workers and others who help Americans in precarious situations.

The funders announced Thursday that they will create a new entity, NextLadder Ventures, to offer grants and investments to nonprofits and for-profits to develop tools for those who often manage huge caseloads with few resources.

“The solutions that we’re investing in, the hundreds of entrepreneurs that are going to bring forward solutions that incorporate leading edge technologies, are going to do it by coming alongside people who are living through some of the struggles in the economy,” said Brian Hooks, CEO of Stand Together, a nonprofit started by Kansas-based billionaire Charles Koch.

The other funders include hedge fund founder John Overdeck and Valhalla Foundation, which was started by Inuit cofounder Scott Cook and his wife Signe Ostby. Ballmer Group is the philanthropy of former Microsoft CEO Steve Ballmer and his wife Connie. The funders declined to reveal the exact financial commitments made by each of the contributors.

The point of investing in these AI tools is to spur economic mobility, a focus all the funders share, they said. The funders believe there are many ideas for how AI technologies could help match people with resources after a disaster or an eviction, for example, or help a parole officer close out more cases for people who have met all of the criteria but are waiting for the paperwork to be processed.

“As we traded notes on where we were making investments and where we saw broader gaps in the sector, it was readily apparent that there was a real opportunity to come together as a group of cofunders and cofounders to establish a new kind of investment organization,” said Kevin Bromer, who leads the technology and data strategy at Ballmer Group. He will also serve as a member on NextLadder's board, which will include three independent board members and representatives from the other funders.

NextLadder will be led by Ryan Rippel, who previously directed the Gates Foundation's economic mobility portfolio. The funder group has not yet determined if NextLadder will incorporate as a nonprofit or a for profit organization but said any returns they make from investments will go back into funding new initiatives.

Jim Fruchterman, founder of Tech Matters and author of the recent book “Technology for Good,” said he expects NextLadder to mostly fund nonprofits if they want to accomplish their mission of reaching the poorest people and places. He said he was optimistic about their focus on serving frontline workers rather than trying to replace them.

“The nonprofit sector is about humans helping humans,” Fruchterman said. "And if instead of inflicting the AI on poor people, or people in need, we’re saying, 'Hey, you’re a frontline worker. What’s the crappiest part of your job that is the least productive?' And they’ll tell you and if you work on that, you are likely to be more successful.”

NextLadder will partner with AI company Anthropic, which will offer technical expertise and access to its technologies to the nonprofits and companies it invests in. Anthropic has committed around $1.5 million annually to the partnership, said Elizabeth Kelly, its head of beneficial deployments, which is a team that focuses on giving back to society.

“We want to hand-hold grantees through their use of Claude with the same care and commitment we provide to our largest enterprise customers,” Kelly said, referencing Anthropic's large language model.

Hooks, of Stand Together, said philanthropy can reduce the riskiness of these types of investments and offer organizations more time to prove out their ideas.

“If we’re successful, this will be the first capital to demonstrate what’s possible,” Hooks said.

Suzy Madigan, who is the Responsible AI Lead at Care International UK, has researched the risks and benefits of using AI tools in humanitarian contexts. She said she's seen a rush to explore how AI technologies might fill in gaps as funding has been cut.

“The rise of artificial intelligence being deployed in more sensitive contexts brings some really important new ethical and governance questions because it can actually exacerbate increasing inequalities, even when there were good intentions behind it," said Madigan.

The key to not harming vulnerable communities is to involve them in every step of developing, deploying and assessing AI tools and to ensure that those tools do not replace frontline workers, she said.

Researchers like those at the Active Learning Network for Accountability and Performance in humanitarian action have studied some of the risks associated with using AI tools when interacting with sensitive populations or handling high-stakes interactions, for example, in humanitarian contexts.

They recommend assessing whether AI is the best tool to solve the problem and, crucially, if it works reliably and accurately enough in high-risk settings. They also recommend assessing tools for bias, considering privacy protections and weighing the cost of potential dependence on a specific provider.

The National Institute of Standards and Technology also emphasizes that trustworthy AI systems should be accountable to users and that it should be possible to explain or trace how a tool arrived at a certain conclusion or decision.

Hooks emphasized that any AI tools NextLadder invests in will be shaped by the needs and feedback of these frontline workers. Tools that don't work for them, won't succeed, he said. Even with the potential risks of AI tools, he said it's imperative that groups that are struggling to move up the economic ladder have access to new technologies.

“The idea that we would deprive those who are struggling in our country from the benefits of the leading edge solutions is unacceptable,” Hooks said.

___

This story has been updated to correct the name of Scott Cook.

___

Associated Press coverage of philanthropy and nonprofits receives support through the AP’s collaboration with The Conversation US, with funding from Lilly Endowment Inc. The AP is solely responsible for this content. For all of AP’s philanthropy coverage, visit https://apnews.com/hub/philanthropy.