Non-profit leaders are struggling to figure out how AI fits into their organizations while navigating ethical, privacy, and governance concerns. Contributor Julie Sobowale looks at some of the challenges and opportunities.
It’s not easy reaching artists who need legal assistance. And because lawyers are expensive, artists usually don’t reach out when they need help.
Martha Rans, supervising lawyer and legal director for the Pacific Legal Education and Outreach Society (PLEO), works hard to reach artists and other non-profits that support artists. The PLEO team created Law for Non-Profits, an interactive website that provides answers to legal questions, but they wanted to enhance it. They decided artificial intelligence was the answer. Rans and her team created a chatbot that would allow artists to write bylaws for creating an artist collective or interpret a clause in a contract. With five contract workers and a $40,000 budget, they finished the project in six months.
Building the chatbot wasn’t easy. In early testing, it was only 60% to 70% accurate when giving responses. They had to go through 90,000 words of legalese, rewriting it into plain language to make it easier for the tool to understand and to improve its accuracy. “It was hard to explain to developers why this wasn’t good enough,” Rans says. “People are using ChatGPT more than [any other AI tools], but it’s not good enough. We’re a legal organization and we have to follow ethical considerations like making sure the information we publish is correct.”
People are using ChatGPT more than [any other AI tools], but it’s not good enough.
Martha Rans, Pacific Legal Education and Outreach Society
PLEO had what non-profits need to succeed in AI: leadership, financial resources, and skilled staff. Non-profit leaders are struggling to figure out how AI will fit into their organizations, and many don’t know where to start. lndustry leaders are working on creating new resources for organizations to make the leap into AI.
Last summer, the Human Feedback Foundation; policy think tank The Dais, based at Toronto Metropolitan University; and the Creative Destruction Lab launched the Responsible AI Adoption for Social Impact (RAISE) pilot program aimed at supporting non-profits in adopting AI. The initiative includes an accelerator program from Creative Destruction Lab to support five non-profits, online courses created by The Dais on AI governance, and an AI tool kit for non-profits.
There aren’t enough tech workers in the non-profit sector. The 2025 report Futureproofing the Community Service Workforce from the Canadian Centre for Nonprofit Digital Resilience (CCNDR) looked at the growing demand for and short supply of tech workers in the non-profit sector. Out of the 2.5 million Canadians working in the sector, less than 1% have jobs focused solely on technology. Wages in non-profits for tech workers are also below average compared to other sectors, meaning organizations need to explore upskilling their existing workers to fill in the gaps. The report also recommends that funders offer more grants toward building digital skills capacity. CCNDR, in partnership with Microsoft, launched the National AI Impact Hub for Nonprofits, where organizations can access free resources about AI governance and take free courses on how to use AI.
The ethical issues around AI remain an ongoing challenge for non-profits and the broader community. There are significant environmental concerns, with AI data centres using large amounts of energy and water. A report from the International Energy Agency found that searching on ChatGPT uses 10 times more electricity than a Google search. Black, Indigenous, and other equity-seeking groups have historically been marginalized in the use of AI. A recent article from The Guardian outlined how international nongovernmental organizations are using AI-generated images of racialized people in poverty as part of marketing campaigns. Funders are struggling to deal with an influx of grant applications generated by AI.
Non-profit leaders will have to figure out whether to use AI tools, and which ones. The major AI suppliers – OpenAI, Microsoft, and Google – are fighting for market dominance, with hopes that Canadian non-profits will use their familiar tools. Recognizing that non-profits have limited budgets, some companies offer discounts, with OpenAI’s ChatGPT offering 25% and Microsoft 365 Copilot at 15%.
Google Canada is taking a different approach. In 2025, it announced a $13-million AI Opportunity Fund to help upskill Canada’s workforce. According to Google’s Economic Impact Report, generative AI has the potential to boost Canada’s economy by $230 billion. “We are committed to making AI accessible,” says Sabrina Geremia, vice president and country managing director at Google Canada. “Canada is uniquely positioned to capture the immense opportunity AI presents, having been a leader in AI research for years. However, we’ve lagged in AI adoption and literacy. Our goal is to reverse that trend.”
Skills for Change created a free two-week AI upskilling course to be accessible for newcomers, racialized communities, women re-entering the workforce, and youth facing barriers to employment. The Alberta Machine Intelligence Institute (Amii) is looking to incorporate AI literacy into Canada’s post-secondary education. The organization received $5 million to create AI curriculum through its AI Workforce Readiness program. The goal is to reach 125,000 post-secondary students and to complete two national studies on AI skills gaps and the future of work.
“We’re seeing an incredible interest in participation,” says Adam Danyleyko, who works on the Amii training team. “With our initial goal of 25 members to launch, we have 29 post-secondary institutions, and we continue to grow.”
Creating a skilled workforce
PLEO had time and resources to train staff on how to use AI, something that most non-profits struggle with. The Dais hopes to change that. Through the RAISE program’s online courses – one for non-profit leaders and one for knowledge workers – the goal is to educate 500 non-profit workers on AI governance. The leadership course covers AI governance issues such as privacy and copyright; the second course is more practical, with students creating AI projects to implement at their organizations. This is the first step in getting Canada’s non-profit workforce trained in AI.
“Canada’s non-profit workforce is large,” says André Côté, executive director of The Dais. “Even though there’s a potential social impact in using AI, non-profits are under-resourced and under-skilled for adoption. We want to equip workers with the skills for responsibly governing and using AI and see what works.”
We need [non-profit leaders] at the table [shaping policy], but if they don’t understand AI and how to use it responsibly, they will struggle to contribute.
André Côté, The Dais think tank
The courses go beyond learning how to use AI. There’s a need for non-profit leaders to understand AI in order to be involved in shaping legislative policy. “We need their voices at the table, but if they don’t understand AI and how to use it responsibly, they will struggle to contribute,” Côté says.
Dan Kershaw wants to learn how to use AI responsibly. In 2014, the former telecommunications executive joined Furniture Bank in Toronto as its new CEO and was tasked with effecting a digital transformation in the organization. He was new to the charitable and non-profit sector and couldn’t find any funding to support digital infrastructure.
Things changed in November 2022 when ChatGPT was launched. Kershaw started experimenting with ChatGPT along with Midjourney and other new AI tools, trying to figure out what the programs could do and how to use them. He encouraged his staff to experiment, too, and they began to use AI in their marketing and communications. “Generative AI is unique because you have to experience it first to really understand it,” Kershaw says. “There’s a fear of the unknown, but to learn AI you have to play with it. You have to start.”
We talk a lot about ethics and bias. We’re creating reasons not to engage instead of [focusing on] the opportunities we have. Nobody’s waiting for non-profits to catch up.
Dan Kershaw, Furniture Bank
Kershaw was curious about how other organizations are managing AI adoption. That’s why he signed up for the 12-month accelerator program run by Creative Destruction Lab at the University of Toronto’s Rotman School of Management. Five organizations meet biweekly to learn about AI trends and to share their experiences in planning and executing a complex AI project. Kershaw’s organization is working on using AI to streamline its furniture donation process. “AI could have a big impact by making things easier for charities to operate,” he says. “I don’t see that discussion. We talk a lot about ethics and bias. We’re creating reasons not to engage instead of [focusing on] the opportunities we have. Nobody’s waiting for non-profits to catch up.”
Furniture Bank uses AI to share stories about its impact in communities. Because of privacy concerns, the organization doesn’t take images or videos of recipients of donated furniture. Staff use testimonials from clients and donors to write prompts to create AI-generated images. They’re experimenting with AI videos and may have an AI “reporter” in the future.
“We need a change in mindset about AI,” Kershaw says. “I don’t want non-profits to be left behind when we talk about AI. It’s not enough to talk about funding and upskilling workers. I want more people to talk about AI in a positive way.”
Kershaw joined the accelerator program because he wanted to have more conversations with others about how to promote AI use in the sector. “The most important thing is for people just to start,” he says. “Start the journey at your own pace.”
CanadaHelps, one of the other organizations in the accelerator, has been moving quickly in AI. CanadaHelps is a tech non-profit, helping donors connect to more than 86,000 charities across the country. Chief operating officer Jane Ricciardelli says CanadaHelps was invited to be a part of the accelerator program because of past experience in AI. The organization is looking to join another AI accelerator next year. “We’ve been using AI, but we didn’t have a strategic framework,” Ricciardelli says. “I like the approach of identifying the problem first and the desired outcomes, then determining if AI is the right solution to the problem.”
CanadaHelps launched its AI chatbot in 2024 to help users with customer support. So far, the chatbot has decreased – significantly – the number of email requests for tax receipts. “We want to automate more tasks with AI,” Ricciardelli says. “We want to do more with the people we have. That way staff are freed up to do higher-value work.”
Why can’t Canadian foundations fund an AI tool that we own? Why are we enabling the tech sector? They’re making a living out of us.
Martha Rans
Even with upskilling programs and support, funding remains the major barrier. Then there are the privacy and governance issues. At PLEO, they use a US provider for the chatbot, which Rans admits is not ideal. All of the major AI providers are outside of Canada, and with little regulation in the United States there are ongoing copyright and privacy issues about how information is being used. “Why can’t Canadian foundations fund an AI tool that we own?” Rans asks. “Why are we enabling the tech sector? They’re making a living out of us. It’s not about just being in the room but controlling the means of production. We’re not building capacity if we don’t control the tech.”
The focus must be on the mission, not on technology. At CanadaHelps, the next project will be using AI to review the source code for its website. Ricciardelli believes the solution is to focus on mission-driven outcomes and tie in AI as a way to further the goals of a non-profit. “The business outcome isn’t reviewing the code but launching new products faster so we can help more charities,” she says. “If we invest in our infrastructure and tie that to the broader impact, we can get our staff, our board, and our donors to be passionate about technology.”
