📨 AI for Social Impact Newsletter

👋 From the Editor

Hi! I’m Joanna. I’m on a mission to help folks in the social impact sector understand, experiment with, and responsibly adopt AI. We don’t have time to waste, but we also can’t get left behind.

Let’s move the sector forward together. 💫

🧠 EDUCATION

• AI Agents (Agentic AI): AI agents are AI tools that are capable of managing complex workflows and independently pursuing goals with minimal human oversight. Agentic AI works through three key steps: planning (identifying a goal and mapping out the steps needed to achieve it); execution (carrying out those steps across systems without constant human input); and adaptation (monitoring progress, adjusting tactics, and learning from outcomes). Unlike chatbots that wait for your next instruction, AI agents take initiative. But we're still in the early days—today's agents can handle short sequences of tasks but often get stuck or need human intervention. 🤖

• Artificial Generalized Intelligence (AGI): AGI is a topic that inspires both excitement and existential dread. AI researchers predict AGI could arrive somewhere between 2027 and 2040. So what is it? AGI refers to AI that can perform any intellectual task a human can do—understanding, learning, and applying knowledge across a wide range of domains with the flexibility to transfer skills between them. Unlike today's “narrow” AI (which excels at specific tasks), AGI would be able to reason, adapt, and solve novel problems without task-specific reprogramming. We're still in the era of “narrow” AI, but understanding the trajectory from today's tools → AGI helps us grasp the magnitude of what's coming. 💻

✨ INSPIRATION

• AI Agents Raising Money for Charity: In an interesting experiment, nonprofit Sage Future tasked four AI models with raising money for charity, giving them the freedom to choose which charity to fundraise for and how to drum up interest. While they weren’t fully autonomous—as the agents could take suggestions from the human spectators watching their progress—they still coordinated with each other in a group chat, sent emails, created and edited Google Docs together, researched charities, and even created an X account for promotion. One agent signed up for a free ChatGPT account, generated three different images, created an online poll to see which image viewers preferred, then downloaded that image and uploaded it to X to use as its profile picture. While the experiment had limitations (donations came mostly from the human spectators monitoring their progress), it demonstrated just how resourceful these systems are becoming. 💰

• Agentic AI ≠ Human Agency: In 2024, the Patrick J. McGovern Foundation invested $73.5 million in charitable contributions toward AI solutions grounded in human values, partnering with 144 nonprofit, academic, and governmental organizations worldwide to center human dignity, equity, and shared purpose. President of the Foundation, Vilas Dhar, drew a critical distinction between AI agents and our agency as humans: "We use 'agentic' and 'agency' sometimes as if they're the same thing, but agency is a human virtue. It's about my ability to express my interest in the world. An agentic system doesn't do that. It maybe goes off and does a task, but it won't do it with compassion or kindness or empathy." As agentic AI becomes more prevalent, remembering that technology should amplify—not replace—human agency is exactly the kind of grounded thinking the sector needs. ️☮️ 🌍

🚀 ACTIVATION

• Deep Research: ChatGPT, Claude, and Perplexity’s deep research functionality uses AI agents to execute a task. So give it a try: pick a complex research question like "What are the most effective program models for [your cause area] based on recent studies?" or "What grant opportunities exist for organizations like ours in [specific geographic area, program area, additional context, etc.]?" Then use Deep Research in ChatGPT or Claude (paid accounts only), or Perplexity (free with limits), and see what kind of results you get. This is agentic AI in action! 🔍

• Agentic AI in Your Browser: For those of you who are more adventurous AI experimenters, Perplexity's Comet browser is now available for free. Comet has an AI assistant built into a sidebar that automatically sees what you see. You can ask questions about web pages, YouTube videos, or even Google Docs you're working on without opening a new window. In your workflow, you could ask it to "summarize funder eligibility requirements" or "compare this grant opportunity to the last one I researched"… or check out Perplexity’s resource hub on other ways to use it. ☄️

  Semi-related, check out this article about Amazon’s cease and desist letter to Perplexity, requesting it to stop allowing people to use AI agents to buy things… “Welcome to The Battle for the Future of Commerce”.

🤖 What I Asked AI This Week

How does the concept of philosophy change when we consider our future and artificial intelligence?

When we consider AI and the future, philosophy stops being purely about human experience and must grapple with other forms of intelligence. Classic questions about consciousness, knowledge, and ethics get reframed in disorienting ways - like whether non-biological systems can truly experience or know things. Entirely new philosophical domains emerge around the moral status of artificial minds, and philosophy itself might eventually become a collaboration between different types of intelligence rather than an exclusively human endeavor. 🤔

💬 What Do You Think?

This newsletter is meant for you, so I’d love to hear what you think. 💌 Reply anytime!

👀 ICYMI

If you’re new here, welcome! You can check out the archive of past issues here.