Dataset Collection Tool

Read full text Summarise with VIRAT
Authors: Pinran Gao
Cited by: 0

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

We first build up search terms, consisting of generic AI words such as “AI”, “LLM”, as well as model names “GPT”, “LLAMA”. Then it combines it with ai functions such as “assistant”. There is also a list of set terms such as “ai tools”, “ai study”. Overall, it performs 327 searches in total. During the filtering process, apps with no rating are being neglected. Then there are four layers of filtering. The first layer checks for broad AI indicators, such as “AI,” “artificial intelligence,” “LLM,” or “generative AI.” The second layer looks for specific model names, platform names, or vendor references, such as “GPT,” “ChatGPT,” “Claude,” “Gemini,” “OpenAI,” or “Anthropic.” The third layer focuses on AI functionality terms, including phrases tied to concrete use cases, such as AI writing, AI image generation, AI photo editing. The fourth layer uses highly explicit and strongly relevant phrases, such as “AI image generator,” “AI voice chat,” or “AI writing assistant”. There are 38520results without deduplication, and 5619 unique apps are kept.

Fields of Study & Subjects

Information Retrieval and Search Behavior Artificial Intelligence Applications Topic Modeling

Publication details

Journal / source
Figshare
Publication date
25 March 2026
Language
EN