The integration of artificial intelligence (AI) into human decision-making and learning processes, known as hybrid intelligence, has gained considerable attention. While much research focuses on AI’s role in augmenting human task performance, the concept of “machine mentoring,” where AI serves as an active agent in human skill development, remains underexplored. Machine mentoring leverages AI systems’ ability to transfer knowledge through interactions, enabling less experienced users to internalize expertise and refine decision-making strategies. However, this process also carries risks, particularly the transmission of biases embedded in AI models, which can perpetuate flawed cognitive frameworks in users. This paper identifies a significant research gap in understanding the mechanisms of AI-to-human knowledge transfer, examining the dual potential of machine mentoring to enhance learning or propagate biases. By proposing design principles and ethical considerations, we aim to foster the development of AI systems that promote effective and equitable mentorship. Ultimately, this research establishes a framework for advancing machine mentoring while addressing its profound implications for human learning, cultural evolution, and cognitive development.