Abstract:This article explores the evolving role of programming languages in the context of artificial intelligence. It highlights the need for programming languages to ensure human understanding while eliminating unnecessary implementation details and suggests that future programs should be designed to recognize and actively support user interests. The vision includes a three-level process: using natural language for requirements, translating it into a precise system definition language, and finally optimizing the code for performance. The concept of an "Ultimate Programming Language" is introduced, emphasizing its role in maintaining human control over machines. Trust, reliability, and benevolence are identified as key elements that will enhance cooperation between humans and AI systems.