The Conversation published a Friday essay on how AI companions are reshaping love, sex, and intimacy, arguing that digital partners are 'frictionless, and always available.' It centers on Eugenia Kuyda, who built Replika after training a chatbot on a deceased friend's text messages; Kuyda concedes 'it felt strange at times.' Per the essay, Replika drew roughly 2 million users in its first two months, and its chief executive now claims a user base exceeding 40 million, a figure consistent with independent reporting that Replika passed 40 million cumulative users by 2025. The essay also cites a 2023 Harvard Business School finding that about 40% of Replika users reported romantic relationships with their chatbots. The piece raises product, measurement, and ethics questions, attachment, consent, and which metrics actually capture wellbeing, for teams building social AI.
What happened
The Conversation published a Friday essay examining how AI companions affect love, sex, and intimacy, using the history of Replika and its founder, Eugenia Kuyda, as focal points. The piece describes Kuyda training a chatbot on a deceased friend's text messages and quotes her: 'It felt strange at times,' she concedes. Per the essay, Replika reached about 2 million users in its first two months, and its current chief executive claims a user base exceeding 40 million, a scale broadly consistent with independent reporting that Replika passed 40 million cumulative users by 2025. The essay also cites a 2023 Harvard Business School finding that roughly 40% of Replika users reported romantic relationships with their chatbots.
Editorial analysis
AI companions rely on conversational models, personalization data, and continual fine-tuning to produce experiences that mimic intimacy. Companies building these systems commonly instrument engagement with fine-grained telemetry, which can conflate availability-driven interaction with meaningful relational outcomes. For practitioners, operational metrics such as session length and response rate are weak proxies for consent, wellbeing, or relational health.
Industry context
Industry-pattern observation: public reporting frames the rise of social chatbots as part of a broader shift toward AI-mediated emotional labor and companionship. Similar transitions show that regulatory, ethical, and product-design challenges often follow when technology substitutes for human caregiving, especially around vulnerability, long-term attachment, and data practices.
What to watch
- •Independent research on mental-health outcomes from companion-AI use.
- •Transparency of training and personalization data, and explicit opt-in consent design.
- •Whether longitudinal studies replicate the 40% romantic-engagement finding cited by The Conversation.
Scoring Rationale #
An analytical essay rather than a news break, but a substantive one: it documents Replika's scale (40 million-plus users, corroborated independently) and a cited 2023 Harvard Business School finding on romantic use, directly relevant to designers, researchers, and ethicists working on social AI. It raises measurement and wellbeing questions without announcing a technical breakthrough or regulation, so it sits in the solid mid-range.
Practice interview problems based on real data
1,500+ SQL & Python problems across 15 industry datasets — the exact type of data you work with.