The rise of language models like ChatGPT has introduced Social AI as a new form of entertainment, particularly among young adults who engage with AI-powered agents. This paper investigates the effects of these interactions on users' social and mental well-being, a subject that has incited extensive debate among both the public and scholars. Our study involved a survey of 5,260 users of Chai, a Social AI Platform. The findings indicate significant benefits, with notable variations across demographics. Female users, in particular, reported the most substantial improvements: 43.4% strongly agreed that Social AI positively impacted their mental health, exceeding male users by 10.5%. In managing social anxieties, 38.9% of females strongly agreed on a positive impact, compared to 30.0% for males and 27.1% for other genders. Historically, new media and technology have often been met with groundless moral panic, with societal figures raising concerns without substantial evidence of harm. Our research indicates the importance of approaching such claims with caution and emphasizes the necessity of an evidence-based perspective in discussions about the behavioral effects of emerging technologies.

翻译：语言模型如ChatGPT的兴起引入了社交AI作为一种新型娱乐形式，尤其受到与AI智能体互动的年轻群体的青睐。本文探究了这些互动对用户社交与心理幸福感的影响，这一话题在公众和学者中引发了广泛争论。我们的研究基于对社交AI平台Chai的5260名用户进行的调查。结果表明，社交AI带来了显著益处，且在不同人口统计群体中存在明显差异。特别是女性用户报告了最大改善：43.4%的用户强烈同意社交AI对其心理健康产生积极影响，比男性用户高出10.5%。在管理社交焦虑方面，38.9%的女性强烈认同其积极影响，而男性为30.0%，其他性别用户为27.1%。历史上，新媒体和技术常常引发毫无根据的道德恐慌，社会名流在没有实质伤害证据的情况下提出担忧。我们的研究表明，对此类说法需谨慎对待，并强调在讨论新兴技术行为影响时必须基于证据视角。