In user experience design and marketing research, practitioners often rely on proxy users, or personas [1, 3, 6], when direct access to actual users is limited. These personas are fictional but realistic representations of user segments that help inform design and decision-making processes. However, the emergence of generative artificial intelligence (GenAI) presents a novel dilemma: the ability to create highly convincing user personas without grounding them in actual customer data. This presentation examines the implications of using large language models (LLMs) and other GenAI tools in persona creation. While these technologies offer efficiency gains through rapid and cost-effective persona generation, they also raise concerns about the validity and ethical implications of synthetic user representations. The central dilemma lies in balancing the practical benefits of AI-generated personas against the risk of perpetuating biases, stereotypes, or hallucinated user characteristics that may mislead decision-makers [2, 7, 8]. We propose Retrieval Augmented Generation (RAG) [4] as a general approach to this dilemma. In the ideal situation, GenAI capabilities leverage real user data owned by organizations toward the creation of more accurate and reliable persona descriptions, adhering to the best practices of data-driven personas [3, 5].  The presentation also explores the evolution of persona automation, from manual creation and use (2000-2014), through automated creation with manual use (2015-2024), to potential fully automated creation and use (2025 onwards). We discuss the implications of this progression for user-centered design and marketing automation. The presentation concludes by raising important questions about the future of persona systems, including the potential specialization of language models for specific persona-related tasks and the evolving nature of persona interfaces in an AI-driven landscape. Our work contributes to the ongoing discourse on the integration of AI in user experience research and marketing, offering practical insights for practitioners and emphasizing critical considerations for the ethical implementation of AI-generated personas.