Event Recap

In today’s rapidly-evolving technological landscape, the increasing prominence of AI raises questions about its impact on society. This discussion explored how AI is currently being used in social impact work, as well as the potential opportunities and risks associated with AI adoption. Our speakers shed light on some of the ethical considerations and how we can leverage AI for social good and offered practical tips for getting started with AI. The goal of this event was to equip our community with the knowledge and tools to harness the transformative power of AI responsibly and in ways that align with their organization’s mission and goals and enable them to drive greater impact in the communities they serve.

Trends, Positive Use Cases, and Social Benefits of AI

• Levels the playing field by empowering individuals and organizations to communicate effectively — For individuals who are non-native English speakers, neurodiverse, or tend to use gendered language - companies like Grammarly provide a tool to bridge the language and communication gap, helping everyone become effective communicators, have their intended tone, and feel confident in their communication.

• Provides real-time feedback for students and assists teachers with curriculum development — AI tools can provide a learning experience tailored to individual students’ needs and provide feedback in real-time, enabling them to learn at their own pace. AI can also be used in curriculum development — code.org, for example, has been exploring how AI can assist teachers in assessing students’ work and developing the appropriate curriculum to address students’ knowledge gaps.

• Promotes organizational productivity and efficiency — For organizations that might be understaffed and under-resourced, AI tools can be used for things like prospecting, drafting proposals, grant writing, and more.

Using AI Thoughtfully and Responsibly

• Consider the datasets the AI is trained on — Some datasets might omit important data, be heavily reliant on datasets with little demographic variation, or simply may not be up to date — these are all things that can affect AI output bias. It’s crucial to understand the data fueling your AI tool to effectively serve your community or user/customer base. User research and community engagement can be effective ways to elicit feedback and incorporate lived experiences to develop human-centered AI tools.

• Consider who built the AI and how it was trained — Being able to trace who built an AI tool, how the model was trained, and how the AI arrives at its decisions (output) are all important considerations to understanding the integrity and reliability of the AI tool. Responsible use of AI means ensuring that your AI tools are built with ethical considerations top of mind, like safety, privacy, and bias.

• For organizations deploying or leveraging AI in your services, measure the outcomes and impact of your AI — It’s important to build capacity to understand how your AI performs across different groups and the long term-impacts on these groups. It’s also important to continue measuring top-level outcomes— explainability and transparency are key.

• For builders of AI tools, align incentives between your business and customer/user base — never sell user data for profit, commit to only making money by building a product people and businesses love to use/find valuable and make values-aligned decisions. It is also important to provide mechanisms for users to report issues, and have a rigorous operational process in place to quickly and consistently address problems and improve your models.

Getting Started with AI

• Ask yourself “How can AI help advance my mission and drive more impact in my line of work?” — This can mean many different things, but it will help inform your decisions about what type of tool(s) to build or adopt without blindly pursuing those that might be popular but antithetical to your mission.

• Ensure the AI services you're considering are built responsibly— Ask AI service providers how they prioritize responsibility, safety, and security when building their products and look for AI products that are built to augment your teams, and companies with enterprise-grade certifications and attestations, such as ISO, SOC 2, and HIPAA and GDPR compliance.

• AI tools can be applied across most business functions — AI can serve as a useful tool for people working across marketing, engineering, development, and many other business functions to cut down time spent on tasks such as editing blog posts and press releases, crafting donor strategies, and building out software development processes. “It’s never going to be 100% ready for use, but may save you time 60% of the time so you can focus on doing something else more meaningful.”