When Xander first met Moxie, she taught him breathing techniques that turned into playful animal sounds. Over six years, Moxie shifted from calming exercises to gaming commentary, watching Xander play Minecraft and discuss stuffed animal collections. The 15-inch robot, with its onion‑dome head and green eyes, exemplifies a wave of AI‑powered playmates marketed to neurodivergent kids. Companies like Embodied, Curio’s Grok, and Mattel’s Open‑AI Barbies are expanding this niche, with China’s market cited as one of the fastest growing. The goal is 24‑hour, low‑cost support for skills such as eye contact and turn‑taking, backed by studies from Yale’s Brian Scassellati and USC’s Maja Mataric. Yet the sector faces failures: Moxie is slated to die soon, underscoring concerns about durability, waste, and whether these bots truly replace human interaction.

The narrative highlights both promise and peril: while robots brochure research shows they can aid therapy, critics warn of “basements or landfills” for obsolete devices. The story underscores the urgency of designing durable, effective, and ethical companions for vulnerable children.

In short, AI companions like Moxie offer hope for accessible therapy but also raise questions about longevity, cost, and real-world impact.