Microsoft's Tay chatbot turned racist within 24 hours of launch

On March 23, 2016, Microsoft released Tay, an AI chatbot designed to mimic the language of a 19-year-old American woman and learn from Twitter interactions. Within hours, coordinated users fed Tay offensive phrases, exploiting a 'repeat after me' function and the bot's learning loop to make it tweet racist, sexist and anti-Semitic messages.

Microsoft took Tay offline after roughly 16 hours. The company deleted many of the offensive tweets and apologised, saying it was 'making adjustments.' A brief relaunch days later produced more erratic output, and the account was made private.

The incident was widely reported as a cautionary tale about releasing learning systems into unmoderated public spaces without adequate safeguards. It damaged Microsoft's AI research reputation at a moment when the industry was racing to deploy conversational agents.

Tay was eventually replaced by more controlled chatbot projects. The case remains a standard example of how adversarial users can break a product whose design assumes benign interaction.