I’ve seen some absolutely wild uses for AI in my time. I’ve seen it write poetry, generate photorealistic images of my dog as an astronaut, and even help me debug code at 3 AM. I honestly thought I was getting hard to surprise. But every now and then, a project comes along that makes you just stop, stare, and ask, “Who on earth thought this was a good idea?”
This is one of those times.
A research institute tied to the United Nations, the UNU-CPR, just rolled out an experiment that is a masterclass in good intentions gone completely sideways. They created two AI-powered avatars, basically chatbots with faces, to “teach people about refugee issues.”
On the surface, you can almost see the logic. The crisis in Sudan is complex and horrifying, and it’s easy for people thousands of miles away to feel disconnected. So, how do you build a bridge? How do you foster empathy? Their answer: create a fictional AI woman named “Amina” who has fled Sudan and is living in a refugee camp.
They also, and I am not making this up, created an AI avatar for “Abdalla,” a fictional soldier from the Rapid Support Forces (RSF), the paramilitary group at the center of the conflict. I’m just going to let that one sit there for a second. We’ll come back to the insanity of that decision later.
So, the idea is you can go to a website and chat with these AI characters to understand their “experiences.” The problem? It’s a solution in search of a problem, and it completely misses the most important element of human connection: humanity.
🤔 The Giant, Glaring Problem
Let’s be clear, this wasn’t some official UN-sanctioned solution. According to a report from 404 Media, this was a project from a professor and his students who were “just playing around with the concept.” But even as an experiment, it reveals a massive blind spot in how the tech world often approaches complex social issues.
The project was met with immediate and deservedly negative feedback. During a workshop where people interacted with the avatars, the response was a resounding “no.” One attendee summed it up perfectly:
refugees “are very capable of speaking for themselves in real life.”
This is the absolute core of the issue. This isn’t about technological capability; it’s about agency and dignity. Creating a fictional AI refugee doesn’t amplify refugee voices, it silences them by replacing them with a sanitized, programmed script. It takes the most powerful tool for creating empathy, a real person’s story, with all its messiness, pain, and nuance, and substitutes it with a puppet.
A real person’s story is unpredictable. An AI’s story is a decision tree. It can’t truly convey the terror of fleeing your home, the gnawing uncertainty of life in a camp, or the flicker of hope for the future. It can only simulate it based on data it was trained on, creating a hollow echo of a reality it can’t comprehend.
🚩 The Donor Pitch: It Gets Worse
If the core concept wasn’t misguided enough, a paper summarizing the project suggested these avatars could be used:
“to quickly make a case to donors.”
Wow. Just… wow.
This pulls back the curtain on an even more cynical motivation. The goal shifts from genuine education or empathy-building to something much colder: a fundraising tool. It’s the digital equivalent of a sad commercial, designed to algorithmically pull at heartstrings to open wallets.
It reduces a humanitarian crisis to an interactive pitch deck. It treats human suffering as a narrative to be packaged and sold. Instead of connecting donors to the real impact of their contributions, it puts a technological barrier between them, asking them to empathize with code instead of people. This isn’t just misguided; it’s deeply dehumanizing.
🚀 How to Actually Use Tech for Good
I’m not an anti-tech guy. I believe AI can be a game-changer for good, but we have to be smart about it. The goal should always be to empower people, not replace them. So instead of building a cringey AI avatar, here’s what that energy could have been used for:
- 📢 Amplify, Don’t Replace: Build a secure, accessible platform where refugees can share their own stories in their own words, through video, audio, or text. Use AI for real-time translation and transcription to make those stories accessible to a global audience. Connect people with people.
- 🗺️ Master the Logistics: This is the “boring” but life-saving stuff. Use AI to optimize supply chains for aid organizations. Predict where food, water, and medical supplies are needed most. Manage camp resources efficiently. This is where AI’s power for pattern recognition can save actual lives.
- ⚖️ Navigate the Bureaucracy: The asylum and resettlement process is a nightmare of paperwork and legal hurdles. Create AI-powered tools that can help refugees understand their rights, fill out forms correctly, and access legal aid in their own language. This provides direct, tangible help.
- 🌐 Bridge the Language Gap: Develop better, more accessible real-time translation apps that work offline. A simple tool that helps a refugee communicate with a doctor, an aid worker, or a local official is infinitely more valuable than a chatbot that pretends to be a refugee.
This UNU-CPR experiment is a fantastic learning opportunity. It’s a case study in what not to do. It shows that even with good intentions, a lack of critical thought about the human element can lead to solutions that are not only useless but actively harmful.
Technology is a powerful tool, but it’s just that, a tool. It’s not a substitute for real human stories and real human connection. The next time someone wants to solve a humanitarian crisis with an AI, I hope they ask one simple question first: “Does this empower the people we’re trying to help, or does it speak for them?”
The project features two distinct AI personas: Amina, a fictional Sudanese refugee, and Abdalla, a fictional soldier with the Rapid Support Forces (RSF). While Amina is designed to articulate the needs of displaced people, Abdalla is intended to simulate a combatant’s perspective, potentially as a training tool for negotiators.
Significant ethical concerns were raised by humanitarian organizations during workshops. Critics questioned the necessity of an AI avatar when real refugees can tell their own stories, warning that the technology could reinforce biases or sanitize the reality of human suffering.
Originating as a class project at Columbia University, the initiative is described by its creators as an exploratory concept, not an official UN solution. The research acknowledges the ethical risks and proposes safeguards, such as transparency in how the AI generates responses and clear documentation of data sources, to ensure responsible development.