Navigating the AI Frontier: Why Non-Profits Must Get Artificial Intelligence Right
The Critical Juncture of AI Adoption for Non-Profits
Artificial intelligence (AI) is rapidly evolving, presenting both unprecedented opportunities and significant challenges for the non-profit sector. As organizations grapple with the potential of AI to enhance their operations and expand their reach, a crucial understanding is emerging: the imperative to adopt this technology strategically and responsibly. Recent research, including a new study from the University of Melbourne and KPMG, sheds light on the complexities of AI integration, particularly concerning public trust and risk management within the non-profit landscape. The findings suggest that while public trust in AI is generally low due to concerns about misinformation, job displacement, and data security, there is a higher propensity for trust when AI is deployed with a clearly benevolent purpose—a characteristic inherent to the mission of most non-profits.
Leveraging Benevolence to Build Trust in AI
The core mission of non-profit organizations is often rooted in doing good, a principle that can be a powerful asset in fostering public trust in AI. As Professor Nicole Gillespie of the University of Melbourne noted, "I think people lean in a lot more, and are a lot more forgiving when the whole purpose of the AI is to do good." This inherent alignment between AI’s potential for benevolent application and the non-profit sector’s raison d’être provides a unique advantage. By emphasizing the positive societal impact and ethical considerations in their AI deployments, non-profits can more effectively gain stakeholder confidence. This approach moves beyond mere technological adoption to a more nuanced strategy that prioritizes human values and societal benefit, thereby mitigating the inherent skepticism surrounding AI.
A Risk-Stratified Approach to AI Implementation
A central theme emerging from discussions and research is the necessity of a "risk-stratified approach" to AI implementation. This involves classifying AI applications based on their potential for harm, allowing for a more tailored and cautious deployment strategy. Low-risk AI applications, such as those automating administrative tasks or analyzing non-sensitive data, can be trialled and adopted relatively quickly. Conversely, high-risk applications, which might involve sensitive client data or critical decision-making processes, require a more deliberate approach. This includes implementing additional governance structures, ensuring rigorous testing, and critically, maintaining a "human in the loop" to oversee and validate AI-driven decisions. Emma Crichton, APAC CEO of AutogenAI, aptly summarized this by stating, "AI for everyone is AI for no one," advocating for AI to be used as a "surgical tool to solve a specific problem, not as a blunt instrument applied universally." This highlights the danger of a one-size-fits-all strategy and underscores the need for bespoke AI solutions that address distinct organizational needs.
Good Shepherd’s Proactive AI Journey
The practical application of these principles is already being explored by organizations like Good Shepherd, a prominent non-profit providing services across 15 family violence refuges, operating a 24-hour crisis hotline, and managing extensive financial wellbeing programs. Stella Avramopoulos, CEO of Good Shepherd, shared her organization
AI Summary
The integration of artificial intelligence (AI) into the non-profit sector is no longer a distant possibility but a present reality, carrying both profound opportunities and significant risks. Recent research, including insights from the University of Melbourne and KPMG, underscores that public trust in AI remains low, largely due to concerns about misinformation, job displacement, and data security. However, this research also highlights a crucial finding: people are more inclined to trust AI when its purpose is explicitly benevolent. This principle is a powerful lever for non-profits, whose core mission is inherently altruistic. To harness AI effectively, the sector must adopt a "risk-stratified approach," categorizing AI applications based on their potential for harm. Low-risk applications can be piloted and implemented swiftly, while high-risk uses demand caution, robust governance, and a "human-in-the-loop" system. The notion that "AI for everyone is AI for no one" emphasizes the need for a surgical, problem-specific application of AI, rather than a one-size-fits-all deployment. Organizations like Good Shepherd are already experimenting with AI tools like Co-pilot to support their extensive operations, which include managing family violence refuges and providing financial wellbeing programs. This proactive approach to innovation is becoming essential, as relying solely on external grants for technological advancement is no longer sufficient; non-profits must leverage their own resources for future investment. Building stakeholder trust in AI requires a deliberate, slow progression with consistent human oversight in critical decision-making processes. A problem-led approach, where AI is employed to address specific, identified issues, is key to fostering this trust. Historically, the non-profit sector