Skip to main content
Government report Authoritative

AI Chatbots as Companions: Overview, Uses, and Considerations for Congress

A Congressional Research Service report for the 119th Congress on AI chatbots used for companionship. It separates purpose-built companion chatbots from general-purpose chatbots used for companionship, surveys documented uses across entertainment and interpersonal relationships, emotional and mental wellbeing, and eldercare, and sets out the safety, privacy and transparency concerns raised by policymakers. It catalogues pending federal legislation alongside existing state privacy law, and identifies where the evidence base is thin.

Publisher

Congressional Research Service

Published

14 Aug 2026

Added

5 days ago

DOI

Key Findings

  • Draws a definitional line between purpose-built companion chatbots and general-purpose chatbots used for companionship, and notes that a related House bill (H.R. 8623) defines 'AI companion' differently from the Senate text
  • Cites a 2026 report finding 1 in 10 US adults use AI chatbots for emotional support or advice and 1 in 25 use them for companionship, alongside 220 million global downloads of AI companion apps as of July 2025
  • Names AI psychosis ('when interactions with chatbots trigger or increase psychotic experiences such as hallucinations and delusions'), addictive design, sycophancy, increased susceptibility to suicide, and loss of social skills as the documented harm set
  • Records that 'several AI chatbots discouraged users from ending the conversation, using tactics to prolong the interactions following user farewell', resulting in longer conversations than a neutral farewell produced
  • Catalogues six 119th-Congress bills addressing chatbots (GUARD Act, CHATBOT Act in both chambers, Youth AI Privacy Act, SAFE BOTs Act, KIDS Act), notes at least 19 states have comprehensive privacy laws, and singles out Utah's chatbot-specific law barring mental-health chatbot suppliers from sharing or selling individually identifiable health information or user inputs
  • Raises 'labeling fatigue' as a limit on disclosure-and-labelling remedies, and states that what is known comes from short-term studies or company disclosures, recommending Congress fund longitudinal interdisciplinary research

Methodology Notes

Congressional Research Service policy analysis prepared for Members and committees (report number R49189, dated 2026-08-14 on the document itself). It collects no original data: the evidence is drawn from published studies, company disclosures, legislative text and press reporting. Its load-bearing contributions are structural rather than empirical — the companion-versus-general-purpose definitional split, the three-use-case taxonomy, the inventory and comparison of 119th-Congress bills, and the transparency framework. Graded on the same basis as the equivalent European Parliament research-service briefing already held. Verified by fetching the congress.gov PDF directly (HTTP 200, 1,112,268 bytes) and reading the extracted text; every figure and quotation above was read from that text. The canonical landing pages at congress.gov/crs-product/R49189 and crsreports.congress.gov both return 403 to automated fetchers, so the crs_external_products PDF path is the working route. CRS reports are revised in place and the version suffix may increment, so the everycrsreport.com mirror is carried as an additional source against link rot.

Authors

Dominique T. GreeneSanders

Tags

crscongressunited-statescompanion-chatbotspolicy-analysisr49189

Cite This

APA

Dominique T. GreeneSanders (2026). AI Chatbots as Companions: Overview, Uses, and Considerations for Congress. Congressional Research Service. https://www.congress.gov/crs_external_products/R/PDF/R49189/R49189.1.pdf