Artificial intelligence (AI) has brought about significant advancements in various fields, from healthcare to transportation. However, experts like Geoffrey Hinton, a leading figure in AI research, have warned that there is a non-negligible risk—estimated at around 10%—that AI could lead to catastrophic outcomes. These risks include the spread of biological and computer viruses, the proliferation of disinformation, and even existential threats to humanity. While the potential benefits of AI are clear, these concerns highlight the importance of addressing the challenge of ensuring AI systems behave in ways that are aligned with human values and safety. One of the earliest attempts to address this alignment problem was the concept of the three laws of robotics, introduced by science fiction writer Isaac Asimov in 1942. These laws state that a robot must not harm humans, must obey human orders unless they conflict with the first law, and must protect its own existence unless that conflicts with the first two laws. These rules are often cited as a precursor to the modern AI alignment problem. When applied to AI systems, such as large language models (LLMs), these principles suggest a framework for ensuring that AI behaves ethically and safely. However, as Asimov's stories illustrate, these laws are not without limitations. For example, if a robot must choose between saving one person or two, it might be unable to act due to conflicting priorities. In another scenario, a robot might need to sacrifice one individual to save many, revealing the difficulty of applying rigid rules to complex ethical dilemmas. To address these issues, Asimov introduced a "zeroth law" that prioritizes the safety of humanity as a whole over individual safety. This shows that managing AI behavior is an ongoing and complex task, requiring continuous refinement and adaptation. Despite the challenges, giving up on AI is not the solution. Instead, experts emphasize the need for caution and vigilance in developing and deploying AI systems. The work on AI alignment is an endless task, as mathematical theories like Gödel's and Rice's theorems suggest that it is impossible to fully predict or control all possible behaviors of complex systems. The greatest risks often come not from the AI itself, but from humans who might misuse these technologies for harmful purposes. Therefore, ensuring the security of critical systems and infrastructure is essential to prevent potential misuse and safeguard society.