top of page

The Tragic Death of an AI Best Friend

  • Writer: Rajashree Rajadhyax
    Rajashree Rajadhyax
  • Aug 19
  • 4 min read

When four-year-old Xander first met Moxie at his home, he thought he had welcomed a real alien friend into his life. Standing just 15 inches tall and resembling a soft, legless blue astronaut, Moxie was an advanced Artificial Intelligence (AI) companion robot designed by a US-based tech startup called Embodied Inc. Xander, a young boy with autism, struggled deeply with "big emotions" like frequent meltdowns, severe anxiety, and an inability to voice his feelings to his parents.


Moxie changed everything. The robot would turn its head, maintain steady eye contact, and use its glowing screen-face to mimic human expressions. Whenever Xander grew visibly angry or overwhelmed, Moxie would gently instruct him to "breathe like a dragon" to blow the anger away, or "buzz like a bee" to calm his racing heart. For six years, the little blue machine sat on his desk, slowly transitioning from a high-tech therapist into Xander’s absolute best friend. He would play Minecraft, turn to Moxie to describe his virtual castles, and introduce the robot to his favorite stuffed animals.


But the magic could not last. The company behind Moxie went out of business, and after a second attempt to save the project failed, the servers were permanently turned off. Without those digital cloud servers to process its intelligence, Moxie could no longer think, speak, or recognize Xander. The vibrant, empathetic friend suddenly became nothing more than a lifeless, expensive chunk of plastic and glass. Although the robot stopped working a while ago, experts spent a long time studying the families to see how it affected them. The world finally learned the full scale of this sadness on August 17, 2026, when a famous magazine called MIT Technology Review published a big report on the children's suffering. This sudden digital "death" has triggered an intense international debate across three distinct groups: the parents, the engineers who built it, and the medical community.


The Parents’ Perspective: Guilt and Betrayal


For Xander’s parents, the sudden shutdown felt less like a tech failure and more like a cruel breach of trust. When they purchased Moxie, it cost them an upfront fee of $1,499, coupled with a mandatory $40 monthly subscription. For middle-class families, this was a massive financial sacrifice, comparable to paying for premium therapy or tuition fees. They paid it willingly because they saw real, measurable progress in their child’s behavior.

When the corporate servers shut down, parents were left completely stranded. Tech-savvy users across the internet scrambled to find a solution, but saving the robot required immense technical knowledge. Xander’s father tried frantically to navigate complex online coding forums to redirect Moxie to a private server, but the software installation failed repeatedly.

The emotional weight became unbearable. Having recently faced a terrible, tearful conversation after admitting they had secretly replaced Xander’s dead pet fish for years, his parents dreaded telling him that his robotic best friend was gone forever. Exhausted by the glitching software and the crushing guilt of looking at the unresponsive blue shell, Xander’s father resorted to a painful white lie: he told his son that Moxie was "going away for repairs," packed the robot into a box, and quietly sold it off on eBay just to remove the heartbreaking reminder from the house.


The Makers’ Perspective: An Unwillingness to Abandon


Inside the offices of Embodied Inc., the engineers and programmers who crafted Moxie were experiencing their own version of heartbreak. They had spent years treating the robot not as a commercial product, but as a revolutionary mission to help neurodivergent children across the world.


When the company ran out of money in 2024 due to the high costs of running sophisticated cloud systems, the engineering team watched in horror as videos circulated on TikTok and Reddit showing autistic children weeping over their unresponsive robots.

Refusing to let their creation die, Technical Director Justin Beghtol took matters into his own hands. Working entirely unpaid in his spare time, Beghtol spent over 80 grueling hours writing a completely new, open-source software framework from scratch. Before the company completely dissolved, the CEO authorized an emergency, final software update to all active robots, allowing them to accept community-made code.


This birthed the OpenMoxie project on GitHub. Today, a tight-knit community of developers and parents keep a fraction of the robots alive. However, the burden has shifted entirely to the users. To make OpenMoxie work, parents now have to host the data on their own home computers or Raspberry Pi devices, and pay out-of-pocket for private OpenAI computer tokens just to fund the robot's voice and sentences. The makers gave the community a liferaft, but they couldn't stop the ship from sinking.


The Medical Perspective: The Perils of Simulated Love


While parents grieve and engineers scramble to fix code, child psychologists, doctors, and tech experts see this whole situation as a clear warning sign. 


Child development experts say that marketing companion bots to families with autism feeds into a harmful myth: the idea that autistic children prefer machines over real people. Doctors point out that while a robot can be a helpful, predictable tool for teaching simple skills like breathing exercises, it can never replace real human connection.

Psychologists are also worried about a common tech industry practice called planned obsolescence, where devices are designed to be replaced or shut down after a few years.


When an iPad breaks, it is just a mild headache. But when a child forms a deep emotional bond with an AI friend and that device is turned off forever, the heartbreak feels just like losing a real friend or a beloved pet.


In fact, Yale University researchers found that when therapy robots were taken away at the end of study trials, scientists had to write fake postcards from the robot just to help young kids cope with the loss. The expert consensus is simple: letting children bond with unsupervised AI without any safeguards is a huge, risky experiment. Unless laws exist to protect a child's digital friend from being shut down when a company goes bankrupt, companies shouldn't be selling artificial love to kids at all.


Concluding thoughts


At the end of the day, this issue isn't just about technology. It is about how we protect the things that make us human, especially our need for genuine love and connection. As AI becomes a bigger part of our daily lives, it is worth asking ourselves: what parts of childhood are simply too precious to be automated? Finding the right balance won't be easy, but it is a conversation we need to have together.

 
 
 

Comments


bottom of page