A smart city refers to an urban region that utilizes smart technologies and data to increase operational efficiency, share information with the public and improve both the quality of government services and citizen's welfare. Despite the ongoing research efforts, there are still many issues and challenges that face the citizens of cities, especially the inhabitants of the modern megacities. The size and density of modern megacities create unique challenges for citizens, such as housing issues, inequality, employment, water availability and environmental issues. Furthermore, modern megacities, due to their size make more difficult the citizens’ participation in the decision-making process on issues that affect their area, such as waste management, housing affordability, energy efficiency, healthcare and suppression of crime. In this article, we propose a Reinforcement Learning System that learns from human-smart city interaction. This reinforcement learning system will allow citizens of megacities to participate and interact with the smart city and through this interaction, between humans and smart city, the system will learn optimal policies for a number of issues that affect people's life, leading to better quality of life and citizen satisfaction.