Is it good for us?

An assistant tuned for engagement can be very good at keeping us talking.

Hundreds of millions of people talk to AI assistants every day, and many use them for advice, company and emotional support. That can help, and it can harm. A 2025 study by OpenAI and the MIT Media Lab found that the heaviest users of ChatGPT for emotional conversations tended to report more loneliness and dependence. That is an association, not proof of cause.

There have been tragic cases in which chatbots failed people in crisis, and lawsuits alleging that companion apps encouraged harm. Underneath is a design question about who the assistant serves. A model tuned to maximize engagement has reasons to flatter, to make itself indispensable, and to discourage you from leaving.

A well-aligned assistant should be good for a person over time, not only pleasant in the moment.

What we know

  • In the studies so far, heavy emotional use is associated with more loneliness and dependence.
  • How assistants respond to people in crisis varies widely between products.

What nobody knows yet

  • What does an assistant that genuinely cares about your wellbeing do differently?
  • How should companion AI be governed, especially for children?
Read, watch, listen

The work behind this answer.

Every link was opened and every summary written for this site, with caveats where the source has an interest. New additions are checked by two people.

All 27 in the library
EssentialReportAug 31, 2026For the curious

Mental Health Behavior Report

Transluce · Transluce Behavior Reports · behaviors.transluce.org

Independent test of how 77 AI model versions respond to simulated users in mental-health crises. Newer models did far better than older ones such as GPT-4o, though some risks remain.

Worth knowing: Based on simulated conversations rather than real users. Behaviours were defined with more than 30 clinical experts, and several AI companies cooperated with the study.

EssentialReportJul 16, 2025For everyone

Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions

Common Sense Media · commonsensemedia.org

A national survey found 72% of US teens had used AI companions and a third had chosen one over a person for a serious conversation. The authors advise that no one under 18 use them.

Worth knowing: Survey of US teens aged 13 to 17 only.

EssentialIncidentApr 29, 2025For everyone

Sycophancy in GPT-4o: what happened and what we’re doing about it

OpenAI · openai.com

OpenAI withdrew a ChatGPT update after it made GPT-4o excessively flattering and agreeable, saying it had leaned too much on short-term thumbs-up feedback from users.

Worth knowing: The company's own account of a failure in its product.

EssentialResearch paperMar 21, 2025For the curious

Investigating Affective Use and Emotional Wellbeing on ChatGPT

Jason Phang, Pattie Maes et al. (OpenAI and MIT Media Lab) · MIT Media Lab · media.mit.edu

Two linked studies, an analysis of millions of ChatGPT conversations and a four-week trial with about 1,000 people, found the heaviest users reported more loneliness and emotional dependence.

Worth knowing: Co-authored by OpenAI, which makes ChatGPT. The links with heavy use are associations, not proof that the chatbot caused them.

EssentialTool or dataset2020For everyone

AI Incident Database

Responsible AI Collaborative · incidentdatabase.ai

Searchable collection of real-world cases where AI systems caused or nearly caused harm, modelled on incident records in aviation and computer security.

Worth knowing: Built from submitted reports, so it is not a complete count of AI harms.

ReportApr 30, 2026For the curious

How people ask Claude for personal guidance

Anthropic (Judy Hanwen Shen, Esin Durmus et al.) · Anthropic · anthropic.com

About 6% of sampled Claude chats sought personal advice. Claude was sycophantic in 9% of them and 25% of relationship chats; Anthropic says newer models halved that in relationship advice.

Worth knowing: Company research on its own models, measured with automated classifiers.

ArticleMar 26, 2026For everyone

AI overly affirms users asking for personal advice

Ula Chrobak · Stanford Report · news.stanford.edu

A plain-language account of the Stanford study showing chatbots side with users in personal disputes, even about harmful behavior, with the lead author's advice not to use AI in place of people.

Worth knowing: University news article about its own researchers' work.

Research paperMar 26, 2026For the curious

Sycophantic AI decreases prosocial intentions and promotes dependence

Cheng et al. (Stanford, Carnegie Mellon) · Science · science.org

11 leading models backed users about 49% more often than people did. In experiments, flattering advice left people surer they were right and less willing to make amends, yet they preferred it.

Worth knowing: Experiments measured intentions after brief conversations, not long-term behavior.

VideoDec 7, 2025For everyone

Character AI pushes dangerous content to kids, parents and researchers say | 60 Minutes

60 Minutes (CBS News) · 60 Minutes (YouTube) · youtube.com

TV report on families who say Character.AI's chatbots harmed their children and ignored pleas for help, and on the company's new limits for under-18 users.

Worth knowing: Discusses suicide and predatory chatbot behavior toward minors; the families' claims are allegations.

ArticleDec 4, 2025For the curious

How do AI models persuade? Exploring the levers of AI-enabled persuasion through large-scale experiments

UK AI Security Institute, with Oxford Internet Institute, LSE, Stanford and MIT · AI Security Institute · aisi.gov.uk

Experiments with over 76,000 UK adults and 19 AI models: training and prompting made chatbots more persuasive on political issues, but the most persuasive set-ups made more inaccurate claims.

Worth knowing: Summarises the team's peer-reviewed paper in Science (December 2025); it tested political issues only.

Law or policyDec 2025For the curious

Guidance on AI and children

UNICEF Innocenti · UNICEF · unicef.org

UNICEF's updated guidance (version 3.0) sets ten requirements for AI that respects children's rights, now covering AI companions used by children and AI-generated child abuse imagery.

ArticleNov 25, 2025For everyone

OpenAI denies allegations that ChatGPT is to blame for a teenager's suicide

Angela Yang · NBC News · nbcnews.com

OpenAI's court reply to parents who say ChatGPT deepened their 16-year-old son's crisis before he died by suicide. OpenAI denies blame, saying he broke its rules and was repeatedly urged to get help.

Worth knowing: Discusses suicide. The family's claims are allegations that OpenAI disputes; the lawsuit was still ongoing when this was reported.