(iStock/Alona Horkova)
There’s arguably no group more in need of free AI tools and expert staff embedded to deploy them than the nonprofit sector. While the dissolution of USAID and other funding cuts exacerbate growing global demands to address public health, economic opportunity, and climate-related challenges, NGOs and social impact organizations are underwater, triaging widespread and complex needs against limited capacity.
Sensing that strain, AI juggernauts such as Anthropic and OpenAI, among others, are creating a range of grants and free staffing solutions worth hundreds of millions of dollars to help fill the gaps.
Anthropic's recent announcement of Claude Corps—a $150 million initiative that will provide AI-trained fellows, grants, and AI tools to nonprofit organizations—is one example. In June, OpenAI’s foundation arm announced a new $50 million commitment to its two-year-old People-First AI Fund, which aims to support nonprofits and other public service organizations using AI. And Salesforce and Google.org offer AI accelerators for social impact organizations that include months of training, pro bono access to experts, and funding.
The social sector is getting an incredible offer that almost never happens. But the need for careful assessment is real. The rise of nonprofit-targeted philanthropy from the AI industry raises difficult questions about the risks of long-term organizational dependency on free (at least for now) tools, data stewardship, market influence, and the extent to which a small number of technology providers may shape how the sector adopts and integrates AI over time. How nonprofit organizations respond now can shape both their short-term operational capabilities and their long-term financial sustainability.
The Hidden Risks of AI Industry Philanthropy for Nonprofits
Before AI tools and staffing are deployed across social impact organizations, a deep risk assessment across three categories of inquiry can help to mitigate potential adverse and unexpected outcomes of “AI for good” offerings:
1. Who sets the rules around data stewardship? Nonprofits are uniquely tasked with storing some of the most sensitive data sets across sectors. From health records including data around rape and abuse, disabilities, and homelessness, this information offers a vulnerable and comprehensive accounting of issues that don’t necessarily show up on the open web. When it comes to AI industry support for the social sector, beneficiary organizations should play an equal, if not dominant, role in setting guidelines for how AI companies' tools read and store their data, and whether that data can be used to train AI models.
2. How can we prevent long-term organizational dependencies? Despite advances in technology, many nonprofits know the experience of being stuck with ill-fit or costly systems, such as constituent relationship management (CRM) tools, because of the significant labor and financial investment required to change or update them. AI industry funding offers a potential solution to this problem, but some may rightly worry about finding themselves in the same sticky position a few years down the road. To avoid similar challenges with AI tools, as nonprofits begin to receive gifted technology from a single provider, staff should decide in advance whether and how they can eventually switch platforms more easily and cheaply if needed. Other questions to consider: How can they objectively determine which AI tools and companies work best for their needs? What happens after the free tools and staffing provided to these organizations are no longer available? Proactive planning to address these scenarios is needed to ensure nonprofits can effectively take advantage of these gifts while moving on to other tech solutions, if needed, in the future.
3. Are nonprofits leaving money on the table? What could nonprofits stand to gain by coming to the AI industry table as partners rather than beneficiaries? Does the value of the social sector’s inner workings and data—which AI companies will ultimately benefit from—exceed the value of these gifts? What are alternative options to the ones presented for nonprofits that accept AI gifts, or for the sector at large? Approaching AI industry philanthropy with such questions could result in additional long-term benefits, operationally and financially.
Building Guardrails Before Adverse Effects Take Root
The most comprehensive way to mitigate these risks is to establish federal and sector-wide policies, standards, and best practices for gifting AI tools, infrastructure, and staffing, as well as their use by beneficiary organizations. Any such standards should be informed by people in the nonprofit sector and in the communities they serve. Before writing policies, the sector should convene to address these opportunities strategically.
Some of this is starting to happen. Independent Sector recently announced the creation of a panel of experts to develop an ethical framework for the nonprofit sector’s adoption of AI. And resources such as NTEN’s AI for Nonprofits Resource Hub, developed in partnership with the National Council of Nonprofits and Maryland Nonprofits, offer sample AI governance templates that organizations can use as a starting point.
However, to date, nonprofits have little comprehensive guidance on whether and how to accept AI industry gifts, how to set parameters to avoid being locked into a single provider, and when to negotiate the terms of acceptance. Creating such a framework now for nonprofits can prevent a host of issues in the future.
From AI Access to Financial Sustainability
Given the long-term implications of these new AI industry philanthropic models, it may be time for nonprofit leaders to look beyond what’s offered in these gifts. If AI companies stand to continue their meteoric financial gains, and their donations of tools, time, and expertise are offered to the social sector—for which it may receive data, knowledge capture, and research that couldn’t be gained otherwise—then there’s an argument to be made that nonprofits should also see some of the economic gains. Indeed, approaching these gifts as a starting point rather than a final offer could ultimately help expand the sector’s funding and capacity and sustain its impact at a time when its long-term existence is more precarious than ever.
Expanding the terms of these donations could include options such as:
Equity, stock, and capacity grants: Nonprofits accepting tools and staffing from AI companies might also receive financial assets from the company as part of the philanthropic gift, giving them a stake in the future upside of AI development and contributing to the sector’s overall financial sustainability.
Subsidized nonprofit AI staff: Embedded staff from fellowships such as Claude Corps and similar programs could be converted to permanent hires, continuing their work within an assigned nonprofit after successfully completing their fellowship and allowing the organization to retain that institutional knowledge. Salaries could be subsidized by the host AI company.
Community benefit and mitigation funds: Nonprofits partnering with the AI industry could advocate for ongoing payments to civic institutions working towards proactive efforts to build AI literacy and skills for the community at large, and for those whose jobs might be displaced by AI. Funds can also be used to address any potential harms caused by AI data centers, including potential adverse environmental, physical, and medical impacts on individuals living near them.
This moment represents more than a technological and operational shift for nonprofits. Approaching AI industry gifts cautiously, with an eye toward long-term benefits yet to be realized, could give the sector the opportunity to sustain itself financially and give communities a voice in how AI technology develops and proliferates for social good.
Read more stories by Keosha Varela.