Snap confirms and fixes a My AI glitch that caused the chatbot to post its own Story and then stop responding to users' messages
Snapchat's My AI feature, an in-app AI chatbot launched earlier this year with its fair share of controversy, briefly appeared to have a mind of its own.
Context & Ripple Effects
My AI moved quickly from a Plus-only, pinned chat feature to a free tool with group-chat and recommendation features. That broader rollout was followed by a wave of one-star App Store reviews, making reliability and user control material to the feature's reception.
The chatbot was also becoming strategically important to Snap: the company had begun exploring using My AI conversations for ad personalization. A visible malfunction therefore matters beyond a single support incident because it can affect willingness to engage with the assistant.
First-order effects
- Snap has restored My AI's normal behavior after the incident, while users who encountered the unexplained Story post and unresponsive chat face a direct trust disruption.
- The glitch adds operational and moderation scrutiny to a feature Snap had expanded to all users and group chats, where unexpected activity can be more conspicuous.
Second-order effects
- Product teams must treat public-facing chatbot actions differently from ordinary text errors: failures that appear autonomous can drive support demand and renewed negative feedback faster than a bad answer in a private chat.
- Snap's plans to derive value from chatbot engagement face a constraint: users may limit interaction with My AI if they doubt whether its behavior is predictable, reducing the quality of any downstream personalization signal.
Third-order effects
- As social platforms embed assistants in persistent, highly visible communication surfaces, reliability, clear boundaries on what an assistant can publish, and rapid incident response become product-governance requirements rather than back-end concerns.
- The episode suggests adoption of embedded AI will be shaped not only by usefulness but by whether platforms can preserve users' sense of agency when an assistant is always present in their social space.
The trend: This is one data point in the shift from optional chatbots to ambient social-platform assistants, where trust and behavioral guardrails determine whether distribution translates into sustained use.