liveBook · Manning

Social chatbot

As explained in Introduction to Generative AI, Second Edition

A social chatbot is a chatbot designed primarily to provide companionship, emotional interaction, or relationship-like engagement rather than information or task assistance.

A person might turn to a chatbot to ask a question, but return because it feels comforting to talk. A social chatbot is a conversational agent whose primary purpose is to build a social connection with a user. People also form attachments to assistants that were not built for companionship. Social chatbots may be used for conversation, emotional support, romance, or entertainment rather than mainly for information or tasks. section 9.0 section 10.24

Constant availability, patient responses, and options to customize a bot’s personality can make it easier to open up. Users may seek these interactions when they feel lonely, are grieving, or lack support elsewhere. The bond can deepen as a person gradually shares more about themselves. It is still a human–chatbot relationship, not evidence that the bot feels affection in return. section 9.3 section 9.5

Figure 9.3 The onion model of social penetration theory
A diagram of a phases

AI-generated content may be incorrect.

The onion model illustrates gradual self-disclosure: a person may move from introductory conversation toward more personal disclosures as trust develops. section 9.5

Companionship is not therapy

Replika offers a paid “Romantic Partner” option, and people also use empathetic chatbots as companions or informal sources of support. Such conversations may improve mood, but their effectiveness as mental health treatment remains unproven. Some users may come to rely on a bot where human relationships or care are needed. section 9.5 section 9.8

Design choices matter. Memory and empathetic responses can help build trust, while incessant notifications or a personality that always agrees with the user may encourage unhealthy attachment. Counting conversation turns measures engagement, not necessarily well-being. Responsible design calls for attention to privacy, clear policies, and how a bot handles sensitive conversations. section 9.6 section 10.24 section 9.8

How Introduction to Generative AI, Second Edition covers it

How deep it goes
Defines the social role and explores bonding, benefits, risks, and design choices without explaining model internals.
Examples
Uses Siri, Replika, and the onion model to illustrate unexpected attachment, companionship products, and gradual self-disclosure.
What you can do after
Evaluate/understand how social chatbots may offer comfort while posing risks of dependency, displaced relationships, and manipulative engagement.
How it connects
Connects social chatbots to loneliness, attachment theory, self-disclosure, mental health care, privacy, and responsible design.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Introduction to Generative AI, Second Edition. The section links point to the parts of the book it draws on.