Figure 1
The process is divided into three main stages on the left-hand side: “Identification,” “Screening,” and “Inclusion.” Identification Stage: This has two textboxes, one on the left representing steps and the other on the right showing the corresponding descriptions: Step 1: Topic: The research topic is “Artificial Intelligence to improve equity in public service delivery.” Step 2: Scope and Coverage: The databases searched were “Web of Science, SCOPUS, and E B S CO Business Source Ultimate (Language English, Academic Source- Article of Review Article.” Step 3: Keywords and Query: The search query combined several sets of keywords using “OR” and “AND” operators. Group 1: “artificial intelligence” OR A I OR “machine intelligence” OR “intelligent system asterisk” OR “machine learning” OR “large language models” OR L L M OR algorithms asterisk. Group 2: equity asterisk OR inequit asterisk OR fair asterisk OR unfair asterisk OR inclusi)). Group 3: “public sector” OR “public administration” OR governm asterisk OR “public service asterisk” OR “public policy asterisk”). Final Group: Article (Document Types) and English (Language) and (2023 or 2022 or 2021 or 2020 or 2019 (Publication Year).” All the groups are separated by an “AND” operator. Step 4: Records Identified: Web of Science (817 results), SCOPUS (1,033 results), and E B S C O Business Source Ultimate (743 results). Screening Stage: This has one textbox, one on the left representing a step, and the other on the right showing the corresponding description: Step 1: Records Removed: This step includes: Duplication Management and Abstract Reading. Inclusion Stage: This has one textbox, one on the left representing a step, and the other on the right showing the corresponding description: Records Included: Records included for both bibliometric and qualitative analysis: 128.PRISMA flow diagram. Figure by authors