Artificial intelligence, also known as AI, is a form of technology that enables computers and machines to simulate human intelligence and problem-solving capabilities, and while it is making quite a stir today, AI has been a field of research and development for over half a century.
Understanding the AI Landscape
The term AI was coined in 1956 by a group of scientists who convened for the Dartmouth Summer Research Project on Artificial Intelligence. The convening was organized by a Dartmouth mathematics associate professor, John McCarthy, who believed that every aspect of learning and intelligence could be precisely articulated and simulated by machines. The researchers sought to explore ways to make machines more cognizant, primarily by developing a framework to better understand human intelligence.
Since then, artificial intelligence has made tremendous progress. Unbeknownst to many people, companies have been utilizing artificial intelligence technology in everything from Siri and Google Assistant to Netflix and Google Maps for years. Innovations around generative AI, which focuses not only on analyzing historical data and making future numeric predictions, but also allows computers to produce brand-new outputs that often mirror human-generated content, are currently at the forefront of discussions within the AI field. Businesses and organizations across every sector are working to understand the potential benefits and threats of artificial intelligence.
For the non-profit sector in particular, there is much potential to increase organizational impact through the use of artificial intelligence technologies for things like donor management, stakeholder engagement, service delivery, and operational efficiency. Yet despite the opportunities for positive implementation of AI tools, non-profit organizations face massive challenges to implementation.
Barriers to AI Adoption
In order to mitigate barriers to implementation of AI solutions, non-profits can benefit from considering the underlying reasons for challenges to adoption and exploring best practices from the field for navigating through those challenges.
Financial constraints
Perhaps one of the biggest challenges to the deployment of AI solutions in non-profits is cost. The cost of implementing AI can be significant because it involves data, hardware, software, and specialized personnel. This financial barrier is particularly pronounced for non-profits, which often operate under tight budget constraints than for proft organizations.
With a strategic approach, however, this obstacle can be overcome. For instance, non-profits looking to explore AI solutions can experiment with smaller pilot projects to demonstrate value and scalability before investing in large scale implementation. This could look like selecting a business process that needs improvement, evaluating solutions that meet the requirements of the process improvement and launching a pilot program to evaluate its efficacy. Before making the investment in a full-scale deployment, this method can help nonprofits evaluate the potential advantages of AI solutions without breaking the bank.
Similarly, the best way to save on costs when it comes to deploying a tech solution is to have a clear data strategy.. [add more here and a plug for a data audit maybe?]
Skill Gaps
Lack of information and competence with new AI technologies is another huge barrier to its scalable deployment in non-profit organizations. Without knowledge of machine learning capabilities and/or data security and privacy protocol to adhere to when utilizing machine learning solutions, it can be challenging to decide which technologies are most appropriate for your organization’s needs and how to incorporate them into your workflows.
Investing in training and learning for staff is vital to supporting the successful adoption of AI in company workflows. Upskilling staff can look like formulating a task force dedicated to exploring use cases and usage guidelines, purchasing online learning courses, hiring of learning specialists to facilitate training, or enabling staff participation in conferences and seminars. Increasing your team's access to subject matter experts and providing time for team members to deepen their knowledge can help organizations choose tools that meet their requirements and avoid costly errors that occur when people lack the necessary skills and knowledge to safely use technology.
Data Management and Privacy Concerns
As alluded to above, the adoption of generative AI tools becomes even more complicated when considering issues that arise around data collection, privacy, and security that are associated with the large data models required for the functioning of AI tools. A lack of oversight here can endanger intellectual property and violate consumer privacy, which can have great cost to an organization’s reputability and trustworthiness.
There are several ways to protect against data privacy and security breaches when utilizing AI technologies. They include, as previously mentioned, investing in learning about general data protection regulations, as well as committing to transparency around how user data is used, and implementing robust security measures like data encryption, access controls, and regular security audits before, during and after implementation of AI solutions.
Resistance
Anyone who has ever tried to implement organizational changes knows just how hard it can be to move the needle. Organizations are notoriously slow to change, even in today’s landscape where non-profits often find themselves operating like fast paced startups. There are many reasons for resistance to implementation of new technologies, including fear, misinformation and concerns about job security.
Engaging staff in the AI adoption process through training, task forces and clear communication about AI's benefits is a great way to address skepticism and resistance to AI solutions. Allowing staff time during small-scale projects or pilots to get hands-on, practical experience with AI's applications and benefits can help to build trust and buy-in from all stakeholders.
Ready to begin adopting AI in your organizational workflows?
Download our AI Readiness Assessment, complete with tailored recommendations to enhance AI adoption in your nonprofit today!
Use Cases: How Nonprofits Are Using AI
St. Jude Children's Research Hospital
As part of its digital transformation efforts, this nonprofit’s marketing team utilized Ai solutions to expand it’s donor base in the years following the pandemic. To grow into a new donor base, St. Jude analyzed past donor data to determine that patient stories are the most effective mission communication format and utilized Google’s automated cost-per-acquisition bidding tools to reach new donors with effective language from past campaigns, which resulted in their ability to reach and acquire new donors by 46%.
Crisis Text Line
In 2016, Crisis Text Line, a support hotline that provides 24/7, free, confidential support for people in crisis, all via text, began utilizing an AI algorithm in combination with human engagement (their crisis responders) to analyze online conversations and predict suicide risk. By examining patterns in 2.8 million text messages, their AI model was capable of identifying words most associated with a high risk of suicide, allowing them to prioritize high-risk messages and respond to 94% of them in under 5 minutes. With almost 4,000 text messages received a day this technology is now critical to ensuring that the most at-risk stakeholders receive support during “spike” times of high demand.
Kiva
Kiva is a nonprofit organization crowdfunding loans to unlock capital for the underserved and addressing the underlying barriers to financial access for people around the world. When some loan applications on their platform weren’t getting enough visibility and were going unfunded, Kiva partnered with the DataRobot AI for Good program to improve the way loans are promoted to lenders. They utilized data about each loan application such as daily popularity to predict which loans were unlikely to be funded each day. Once applications with low visibility are identified, the learning model then promotes them on their platform, which helps Kiva more successfully achieve its mission of expanding financial access around the world in underserved communities.
Goodwill
In 2020, Goodwil partnered with Entrupy’s AI-based program to guarantee the authenticity of luxury accessories sold through its auction site and verify designer good in its storefronts. Goodwill combined machine learning algorithms and computer-vision technology to verify items with a 99.1% accuracy rate to support the expansion of its online shopping destination. The automated verification approach means more patrons are willing to pay higher prices for second hand designer items which enables Goodwill to reinvest profits into job training and placement and, community based programs.
As one can see, for nonprofits looking to stay current and improve systems and data usage, it is imperative that leaders recognize and navigate the barriers to AI adoption, in the same way they do for any new change that affects the markets they operate in. Understanding the challenges and implementing targeted strategies to mitigate them is key to successful AI adoption and will help non-profits leveraging AI benefits to further their missions and increase impact.