You are leading a team of junior developers and introducing them to GitHub Copilot. One developer asks if Copilot always provides the best solution. You want to explain the limitations of GitHub Copilot and LLMs, particularly regarding the accuracy and reliability of code suggestions. Which statement best describes a key limitation of GitHub Copilot and other LLMs in generating code?