The development of technology today create a new technology namely Artificial Intelligence (AI). Deepfake as one of the AI has two sides like a knife, it has the positive side is the use of deepfake to support foreign film dubbing, education, and economy (shopping). The negative side is pornography, fraud, and gambling. Indonesia has not yet regulated AI rigidly, so with the legal vacuum of AI in Indonesia, a comprehensive regulation is needed that regulates the liability for AI users against the implications of harmful content by comparing to the United States. This research aims to identify and analyze the urgency of regulating the liability for AI users towards the implications of harmful content; provide recommendations for regulating the liability for AI users towards the implications of harmful content results between Indonesia and the United States. This research is a descriptive normative juridical legal research with a statute approach, comparative approach, and conceptual approach. Sources of legal materials consist of primary, secondary, and tertiary legal materials. Data collection techniques using literature studies were analyzed using interpretation techniques. The results of this study are that with the existence of the National Strategic for Artificial Intelligence (Stranas KA), it is urgent to regulate the liability for AI users for the implications of dangerous content in Indonesia based on the absence of rigid rules governing AI in Indonesia; recommendations for regulating the liability for AI users for the implications of dangerous content in Indonesia must prioritize alignment, emphasis, and a risk-based approach. So that it will realize AI regulation that is based on the ethics and culture of Indonesia.