Abstract: To guard against crises threatening human existence， moral enhancement has entered the agenda of technological ethics. In response to the numerous drawbacks of biomedical moral enhancement， various AI-based moral enhancement approaches have been proposed， which can be categorized as AI serving as a “substitute”， “advisor”， or “interlocutor” in human moral decision-making. The “substitute” approach may lead to the demise of morality and hinder moral progress， but it is suitable for emergencies. The “advisor” approach might diminish human agency and rigidify their moral conceptions， yet it is appropriate for everyday decisions. The “interlocutor” approach might help users achieve moral growth but fails to provide moral motivation. A hybrid moral enhancement approach applies these strategies to different scenarios， invoking the “substitute” mode in emergencies， activating the “advisor” mode in daily moral decision-making contexts， and entering the “interlocutor” mode during the idle time. This hybrid approach capitalizes on the advantages of each approach while avoiding their shortcomings. However， the hybrid approach still requires the support of biomedical moral enhancement to boost users' moral motivation and should be integrated into compulsory educational programs.

Key words: moral progress; moral enhancement; artificial intelligence; existential crisis; biomedical enhancement