Unicorn startup Krutrim, helmed by Ola CEO Bhavish Aggarwal, introduced its AI chatbot in beta mode, but users quickly noticed a concerning issue.
Screenshots circulated showing the chatbot disclosing, "I am an artificial intelligence created by OpenAI."
Krutrim attributed this mishap to a data leakage problem stemming from an open-source dataset utilized in the LLM fine-tuning process. This incident underscores the importance of data security in AI development.