How to encourage LLM users to understand their output before actually using it

Prompt

There is a common pattern developing in which people will use LLMs in order to solve various problems and jump straight to applying or following the solutions without actually properly reading through the generated responses. What techniques would you use to determine them to understand the answers provided by the LLMs firsthand? How can you minimize the chances of the output being used without it being understood by those who generate it? How would you enforce a proper read through the LLM response, respectively a proper understanding of it?