DigestAI news desk
Researchupdated

Study finds trust and friction issues in major generative AI app reviews

Researchers examined 17,012 English‑language reviews from Google Play and the Apple App Store for six leading generative AI applications—ChatGPT, Gemini, Microsoft Copilot, Claude, DeepSeek and Perplexity. Using BERTopic for topic modeling and RoBERTa for sentiment classification, they validated the approach against a stratified sample of 300 human‑coded reviews and applied chi‑square,…

1 source primary source

Key points

  • Analysis of 17,012 English reviews from Google Play and Apple App Store across six GenAI apps.
  • Negative sentiment clusters in advertising (91%), authentication (89%), server reliability (83%) and subscription pricing (73%).
  • Claude shows highest negative sentiment at 47.7% and strong polarization among users.

Negative sentiment was concentrated in four areas: advertising (91%), authentication (89%), server reliability (83%) and subscription pricing (73%). Sentiment varied significantly across applications; Claude recorded the highest negative sentiment at 47.7% while also showing a strongly enthusiastic user base, indicating pronounced polarization. A subset of DeepSeek reviews raised geopolitical and data‑privacy concerns tied to its Chinese origin. The authors introduced a Trust Friction Score to summarize application‑specific trust and usability barriers into interpretable dimensions, offering actionable evidence for developers and policymakers.

Read the original atarXiv cs.CL · by Md Jafrin Hossain, Umme Nusrat Jahan, Shouvaggo Sharif Shammo primary sourceOpen source ↗
Topics · follow one to build your own front page
OpenAIGoogleMicrosoftAnthropicDeepSeekPerplexity AI

The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us.

Comments

via GitHub Discussions

More in Research

All →

Related stories

hot Microsoft exec called AI scraping largest theft of labor in human history, filings show new Is Trump’s AI obsession walking the world into disaster? hot Opinion: Use LLMs only for copyediting, never copy their phrasing, says author new Opinion: Podcast asks if AI deception can be prevented as systems outsmart humans hot Google launches Gemini 3.8 Live and Extended Thinking voice models new New framework optimizes LLM inference costs via adaptive model activation hot Meta CEO Zuckerberg Opposes Coordinated AI Slowdown, Argues Labs Must Self‑Regulate new Qwen3.5-4B outperforms larger LLMs on new user-side conflict benchmark