Ed. note: Throughout the past two years, the Goldhirsh Foundation and its LA2050 initiative have provided complimentary AI (artificial intelligence) consulting, training, and workshops to the Los Angeles impact community and beyond. The work is led by Jen García, Goldhirsh Foundation’s AI EIR (executive-in-residence). Below – and in other posts available in the AI section of our news feed – García shares practical insights, real-world use cases, and emerging research to support nonprofits’ responsible and ethical adoption of AI tools.

This year, our offering of AI literacy training resulted in an overwhelming response by summer. In the beginning of the year, after the devastation we experienced with the two LA fires, some nonprofits in Los Angeles paused or slowed their AI adoption initiatives to focus on supporting our communities. At the same time, nationwide, nonprofits were experiencing an expected but-not-yet prepared-for reduction in federal funding. But after our March AI for Nonprofits training at LA Más in Los Angeles, requests for training resumed nearly three-fold.

Along with training and / or consulting hundreds of people, nonprofits, and small businesses in AI tools, the Goldhirsh Foundation partnered with OpenAI in its very first Nonprofit Jam Day, wherein LA-based nonprofits were trained in everything ChatGPT from prompt engineering to projects to custom GPTs. We also worked with the Los Angeles Fire Department, Brooklyn Public Library, and hosted an LA Tech Week event in partnership with Bloomberg and Fast Forward to support social impact organizations and civil service institutions to leverage AI for good.

But not all nonprofits have been able to catch their breath after such a tumultuous start to the new year, and have not yet begun to learn about the ways in which AI tools can support their work. To assist them in discovering the numerous ways to leverage a generative AI for internal and external purposes, in this blog, we'll discuss easy-to-start applications that any nonprofit can implement today with free generative AI tools.

To review, generative AI uses large language models (LLMs) to generate text. Consumer LLMs are packaged in products such as OpenAI's ChatGPT, Google's Gemini, and Anthropic’s Claude, among many others. Models that generate images or videos are called diffusion models. Both LLMs and diffusion models are included in the umbrella term generative AI.

Generative AI for Research

One of my favorite uses of generative AI is leveraging generative AI tools to support research. Using tools like Perplexity AI or ChatGPT Deep Research is a game-changer for entrepreneurs.

And it doesn't stop with internal processes: generative AI tools can support researching new grant opportunities, establishing policies, and researching best practices from more seasoned nonprofits. In fact, when one nonprofit asked for guidance in creating their AI policy, I showed them how to research AI policy writing best practices using OpenAI’s ChatGPT Deep Research tool, adding to the prompt a request for example policies. Gone are the days of the normal Google search; AI tools can now supercharge all inquiries. As we learned from our earlier blog post interview with Ashley Leon from 9Dots, we can also use ChatGPT to schedule research for us.

Generative AI for Communications

Many businesses and social impact organizations now have the option to use generative AI for various communications, though, we must approach this decision and each task thoughtfully.

Use of generative AI to "write" is contentious for good reason: newly-published research is showing that not only may it be cognitively dangerous for humans to leverage AI tools to offload critical thinking tasks, but when content is determined by the reader or viewer to be AI-generated, trust in its veracity goes down. Understanding this, it is up to the nonprofit’s leadership to steer the AI values of the organization by determining when it is acceptable to leverage these tools.

Some use cases that do not violate my personal AI values are using AI tools for quick marketing content that is merely sharing information, such as an Instagram flyer, an invitation, or a summary of information for an event. Using tools such as Google's new Nano Banana 2 to create infographics for teaching community members, customers, or training volunteers is a useful way to allow AI tools to multiply resources. Due to privacy issues, some nonprofits are also using Nano Banana 2, Midjourney, or other diffusion models to create images of community members being served that otherwise would not be able to be photographed, such as nonprofits helping survivors of domestic violence.

Generative AI for Internal Systems and Documents

Due to the limited resources for those of us in the nonprofit space, the use of generative AI tools to create internal systems and documentation is significant. Now, standard operating procedures and documentation can easily be generated. For example, tools such as Loom can record computer processes, and then Loom's AI-powered standard operating procedure generator to create those steps directly from the transcript of the video. Using this process, nonprofits can record a process once and share it with staff to learn from as many times as needed.

Tools such as Google's Notebook LM, which grounds its results in the sources you provide and then cites exactly where in the source it found its response, are used by nonprofits to house policies that staff searches periodically. This same tool is being used to navigate understanding new laws, regulation changes, and grant opportunities. NotebookLM can also be used by nonprofits that have an educational program within, as it can create customized learning materials such as study guides, podcasts, video presentations, slide decks, and infographics.

The Need for a Generative AI Policy

These simple, out-of-the-box use cases covered in this blog can be implemented today, but it is recommended an organization begin only after a discussion of how to best approach leveraging generative AI. To start, one non-negotiable approach for all generative AI output is to have a person responsible for reviewing and validating the information provided. The term for this person is “human-in-the-loop.” Once that vital practice is understood by all staff members, an easy way to start is to choose the low-hanging fruit of one task or workflow that can be tested for a period of time, document the procedure and tool selected, and begin testing. After said time, leadership should invite a discussion of results and thoughts of those evaluating the AI tools. Then, perhaps a discussion of drafting an organization’s overall AI policy should be in order before more thorough testing is implemented. Some choose to begin their AI policy drafts before even testing tools. Each of these decisions depends on the governance of the organization. No matter how nonprofits choose to proceed, or even if they choose not to proceed, awareness of the capabilities of this tool is important for all of us to understand.

In the first blog of 2026, we’ll share more use cases for nonprofits, along with resources to support your generative AI tools learning journeys.