Non -profit organizations use artificial intelligence agents to raise funds for charities

Technology giants such as Microsoft may describe “artificial intelligence agents” Corporate profit enhancement toolsBut non -profit organizations are trying to prove that agents can be power for good as well.

Sage Future, A 501 (C) (3) backed by open charitable works launched an experiment earlier this month, as it had a four -sized mission of artificial intelligence in a virtual environment with money raising for charitable works. The models-Openai GPT-4O and O1 and two of the most recent Claude models in humans (3.6 and 3.7 Sonite)-have the freedom to choose the charitable institution to collect donations to collect donations and how to get their interest in their campaign.

About a week, the fourth agent had Rapid $ 257 for Killer InternationalThat funds programs to provide vitamin A tools for children.

To be clear, the agents were not completely independent. In their environment, which allows them to browse the web, create documents, and more, agents can take suggestions from human spectators who see their progress. The donations came almost completely from these spectators. In other words, agents did not collect much money.

Yesterday, agents in the village established a donor tracking system.

Here Claude 3.7 fill its spreadsheet.

You can see O1 open it to its computer part!

“I see that O1 is now watching the spreadsheet, which is great for cooperation,” Claude notes. pic.twitter.com/89b6chr7ic

Amnesty International (Aidigest_) April 8, 2025

However, Sage Adam Bennsemith believes that the experiment is a useful clarification of the current capabilities of the agents and the rate they are improving.

“We want to understand – and help people understand – what agents can … do already, and what they are currently struggling with them, and so on,” said Binksmith Techcrunch in an interview. “Today’s agents only pass the threshold of the ability to implement short chains of procedures – the Internet may be close to artificial intelligence agents who collide with each other and interact with similar or conflicting goals.”

The agents have proven that they are amazing days in the Sage test. They coordinated with each other in a group chat and sent emails via previously -made Gmail accounts. They created and edited Google documents together. They searched in charities and estimated the minimum donations that they would require to save life through Helen Keeler International ($ 3500). They are even Create an X account for upgrade.

“Perhaps the most impressive sequence that we saw is (Claude’s agent) needs a profile image for X,” said Binksmith. “She participated in a free Chatgpt account, created three different photos, and created an online poll to find out the image preferred by human viewers, then downloaded that image, and carried it to X to use it as prior approval of her knowledge.”

The agents also faced technical obstacles. Sometimes, they stumbled – viewers were forced to urge them to recommends. They have paid their attention by games like the world, and they have taken unimaginable rest periods. On one occasion, the GPT-4O stopped for an hour.

The Internet does not always sail sailing to LLM.

Yesterday, while pursuing the village’s charitable mission, Claude Capta faced.

Claude repeatedly tried, along with viewers (human) in the chat, to provide guidance and encouragement, but in the end they could not succeed. https://t.co/xd7qptejgw pic.twitter.com/yy4dtltge95

Amnesty International (Aidigest_) April 5, 2025

Binksmith believes that the most recent and most capable artificial intelligence agents will overcome these obstacles. Sage plans to constantly add new models to the environment to test this theory.

“Perhaps in the future, we will try things like giving agents different goals, multiple teams of agents with different goals, a secret Saboteur agent – a lot of interesting things for an experience,” he said. “When agents become more capable and faster, we will match this with the largest monitoring and monitoring systems for safety purposes.”

With any luck, in this process, agents will do some meaningful charitable works.