Cited by 1AI in Service InteractionsArtificial Intelligence in EducationHigher Education Research Studies

Hispanic-Serving Institutions, Community Colleges, and Artificial Intelligence Chatbots: A Systematic Review and Interaction with HSI Community College Websites

Hernández, Susana H., Taylor, Z.W., Ortega, Guillermo·2025·Community College Review

Abstract

Objective/Research Questions : Extending Garcia’s notion of servingness , our study explored the prevalence and usability of artificially intelligent (AI) chatbots within all 234 HSI community colleges according to the most recent HACU list. The research questions for our study are: (a) Do HSI community colleges embed artificial intelligence in the form of chatbots into their websites to facilitate access to information? and (b) To what extent do HSI community college chatbots respond to the needs of Spanish-speaking stakeholders? Methods : We utilized a mixed-methods approach where we explored each institution’s website to locate possible chatbot integration. Once we reviewed each home page, we employed a binary coding strategy (0 = no, no chatbot 1 = yes, chatbot integrated). Then, we interacted with each chatbot—first in English and then in Spanish—during normal business hours to explore their level of interaction and coded these interactions. Results : Findings suggest HSI community colleges rarely utilize AI chatbots, and when they do, they often do not facilitate access to Spanish-language information nor provide the option to chat with a Spanish-speaking human agent. Conclusions/Contributions : Few studies have explored how higher education institutions use chatbots to facilitate access to information, and even fewer studies explore linguistic equity of these AI chatbots. Our findings contribute to research on the use of AI chatbots within community colleges and HSI servingness , specifically demonstrating how providing Spanish and bilingual resources and services through AI chatbots can enhance servingness .

Integrity signals

Registry

Crossref · 10.1177/00915521251338822

Cites

48 of 48 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

4 of 20 cited sources checked · 0 supported · 0 flagged · accuracy not yet measured

  • Interpretationcitation

    The study supports Taylor et al. (2022), who found that it was common for institutions not to have staff who can maintain chatbots and answer live questions.

    “Our study supports Taylor et al. (2022), who found that it was common for institutions not to have staff who can maintain chatbots and answer live questions (Taylor et al., 2022).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Taylor et al., 2022 source

      In “Chatbots Programmed in Spanish (With Varying Levels of Fluency)” (around p. 93), the study reports that 14 live-chat functions received no response within half an hour or had no operator online when tested. That supports a finding of unavailable live assistance, but the paper does not establish that institutions lacked staff to maintain chatbots or that staffing shortages caused the unanswered questions. (LLM, conf=0.85)

      Experimental — accuracy not yet measured
  • Premisecitation

    HSI community colleges also provide access to educational opportunities to other non-native English-speaking communities.

    “HSI community colleges also provide access to educational opportunities to other non-native English-speaking communities (Contreras & Contreras, 2015;Cuellar & Johnson-Ahorlu, 2016).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Contreras & Contreras, 2015 source

      The Abstract says HSIs provide Latinos access to higher education, and the Relevant Literature section says most HSIs are community colleges and that students beyond Latinos may benefit from their support grants. Neither passage establishes that HSI community colleges provide access specifically to other non-native English-speaking communities. Page numbers are unavailable in our copy. (LLM, conf=0.84)

      Experimental — accuracy not yet measured
    • cites Cuellar & Johnson-Ahorlu, 2016 source
  • Findingreasoning

    Live agents at five public HSI community colleges were able to chat in both English and Spanish.

    “Thus, live agents at these five public HSI community colleges were able to chat in both English and Spanish.”
    Quote in the paperSubstantial claim
  • Findingown data

    Only one of the Ocelot chatbots attempted to connect users with a live human agent.

    “Additionally, websites for all of the HSI community colleges with Ocelot chatbots indicated that the chatbot could communicate in at least three languages (English, Spanish, and Chinese), yet only one of their chatbots attempted to connect us with a live human agent.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: The host URL of the chatbot application within the webpage's CSS/programming code reliably identifies the chatbot vendor.
  • Findingown data

    16.9% (11 of 65) of public HSI community colleges referred users to a live human agent who could chat in Spanish.

    “Similar to English, 16.9% (11 of 65) referred us to a live human agent who could chat in Spanish.”
    Quote in the paperSubstantial claim
  • May combine two claims. 20
  • Classified as probably not a substantial claim. 18
  • May combine two claims.

    Premisecitation

    Taylor et al. (2022) found that few institutional websites provided Spanish language admissions information and only 5.7% (19 of 331) of institutions with chatbots could provide information in Spanish, with even fewer chatbots elevating users to a human agent who can communicate in Spanish.

    “They found that few institutional websites provided Spanish language admissions information and only 5.7% (19 of 331) of institutions with chatbots could provide information in Spanish, with even fewer chatbots elevating users to a human agent who can communicate in Spanish (Taylor et al., 2022).”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Taylor et al., 2022 source

      The Results (p. 90) report that few admissions webpages offered non-English content. Table 2 reports 19 Spanish-capable chatbots among 71 institutions with chat functions (26.8%) and only 5 Spanish-fluent live-chat options (7.0%). The claim’s 19 of 331 (5.7%) uses the full institutional sample as its denominator, not the institutions with chatbots, so that qualifier is incorrect. (LLM, conf=0.90)

      Experimental — accuracy not yet measured
  • May combine two claims.

    Premisecitation

    Many AI technology companies were created from a white, Anglocentric perspective, so the tools and features developed for chatbots are often built using English, rather than Spanish meant for Spanish speakers and Latinx people.

    “Research suggests that many AI technology companies were created from a white, Anglocentric perspective, so the tools and features developed for chatbots are often built using English, rather than Spanish meant for Spanish speakers and Latinx people (Taylor & Burnett, 2019;Taylor et al., 2022).”
    Quote in the paperSubstantial claim1 of 2 sources checked
    • cites Taylor & Burnett, 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Taylor et al., 2022 source

      The Major Themes section (around p. 92) describes English-only chatbots and English-language interfaces that limit access for Spanish speakers. The Discussion (p. 95) likewise says few U.S. higher-education institutions programmed their chatbots to speak Spanish. The paper does not establish that many AI companies were created from a white, Anglocentric perspective or that this caused chatbot features to be built in English; its findings concern college admissions chat services. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • May combine two claims.

    Findingown data

    Chatbots at private not-for-profit HSI community colleges could not elevate users to a live human agent who could communicate in either English or Spanish.

    “Chatbots at private not-for-profit HSI community colleges did not respond in Spanish and could not elevate us to a live human agent who could communicate in either language.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    Some chatbots had different language preferences to access website information, but this feature was limiting because chatbot responses neither linked to resources in Spanish nor connected to a live human agent who could communicate in Spanish.

    “some chatbots had different language preferences to access website information; however, this feature was limiting since chatbot responses neither linked us to resources in Spanish nor connected us to a live human agent who could communicate in Spanish.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    Only 29.1% of the HSI community colleges in the sample had chatbots, and a little over half of those chatbots were able to communicate in Spanish.

    “only 29.1% of the HSI community colleges in our sample had chatbots and a little over half of those chatbots were able to communicate in Spanish.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    11 HSI community colleges referred the research team to a live agent who communicated in English, and 11 connected them to a live agent who communicated in Spanish.

    “11 HSI community colleges referred us to a live agent who communicated in English, and 11 connected us to a live agent who communicated in Spanish.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    Chatbots informed users that their services were only available during normal business hours; they did not offer flexible hours.

    “Chatbots would also inform us that their services were only available during normal business hours; that is, they did not offer flexible hours.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    When interacting with live agents, the research team was provided a more personalized experience; live agents tended to introduce themselves and ask questions that seemed less scripted.

    “When interacting with the live agents, we were provided a more personalized experience.Live agents tended to introduce themselves and ask questions that seemed less scripted.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    Few HSI community colleges had chatbots that offered communication in multiple languages; only seven of 67 chatbots provided the option to chat in a language other than English and Spanish.

    “The research team found that few HSI community colleges had chatbots that offered communication in multiple languages.Only seven of 67 chatbots provided the option to chat in a language other than English and Spanish.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    The research team was required to complete forms at seven institutions before they could use their chatbots; these chatbots were associated with five chatbot providers (LibChat, LiveChat, ProvideSupport, SysAid, and Tawk).

    “we were required to complete forms at seven institutions before we could use their chatbots.These chatbots were associated with five chatbot providers (LibChat, LiveChat, ProvideSupport, SysAid, and Tawk).”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    Victoria College provided a personalized experience through a chatbot; its chatbot was responsive to all questions, and when asked to speak with someone, the team was immediately connected to a live human agent, including a Spanish communicating agent.

    “Victoria College was an exemplary HSI community college that provided a personalized experience through a chatbot.The chatbot interaction with Victoria College was responsive to all our questions, and when we asked to speak with someone, we were immediately connected to a live human agent, which ultimately included a Spanish communicating agent.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    Of the 67 HSI community colleges with chatbots, 65 were public institutions and two were private not-for-profit institutions.

    “Of these, 65 were public institutions, and two were private not-for-profit institutions.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    A common practice identified when attempting to communicate with chatbots was the request for personal information; the information requested varied but often included name, email address, phone number, and academic major.

    “A common practice that the research team identified when attempting to communicate with chatbots was the request for personal information.The personal information requested by chatbots varied from one institution to another, but it often included name, email address, phone number, and academic major.”
    Quote in the paperSubstantial claim
  • May combine two claims.

    Findingown data

    The other languages offered by HSI community college chatbots were limited to Chinese and Vietnamese, available only at California institutions.

    “The other languages that were offered were also limited with only Chinese and Vietnamese options, which were only available at California institutions.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisecitation

    Websites help promote an institution's mission and values for students, faculty, administrators, and community members.

    “Websites help promote an institution's mission and values for students, faculty, administrators, and community members (Hutchings & Morgan, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Hutchings & Morgan, 2023

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The study only examined if chatbots could communicate in Spanish.

    “Third, this study only examined if chatbots could communicate in Spanish.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    HSI community colleges with Spanish or bilingual agents are likely more familiar with the needs of the community they serve.

    “These HSI community colleges with Spanish or bilingual agents are likely more familiar with the needs of the community they serve.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Premisecitation

    No study has examined how HSI community colleges leverage AI technology through chatbots and whether it serves Spanish-speaking communities.

    “Yet, no study has examined how HSI community colleges leverage AI technology through chatbots and whether it serves Spanish-speaking communities (Garcia, 2019).”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Garcia, 2019 source
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductnone given

    When communicating with chatbots, the research team only asked questions focused on admissions and transfer information, so access to other information such as housing and student activities was beyond the scope of the study.

    “Lastly, when communicating with chatbots, we only asked questions focused on admissions and transfer information, thus access to other information such as housing and student activities was beyond the scope of our study.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductcitation

    Two research team members were fluent in Spanish at the advanced-mid level according to ACTFL Proficiency Guidelines, though their Spanish fluency might not capture the experiences of all Spanish-speaking individuals interacting with the chatbots.

    “Second, two of the research team members were fluent in Spanish-advanced-mid speakers using the American Council on the Teaching of Foreign Languages (ACTFL) (2012) Proficiency Guidelines-however, we recognize their Spanish fluency might not capture the experiences of all Spanishspeaking individuals interacting with the chatbots.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites ACTFL, 2012

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    According to Garcia (2019), HSIs need to enact servingness through a race-conscious lens and positively affirm the language of Latinx and other minoritized students.

    “According to Garcia (2019), HSIs need to enact servingness through a race-conscious lens and positively affirm the language of Latinx and other minoritized students.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Garcia, 2019 source
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Interpretationreasoning

    HSI community colleges are investing in top AI technology to help serve their students.

    “Though community colleges are often underfunded (Kelchen, 2019), this finding shows how HSI community colleges are investing in top AI technology to help serve their students.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: The host URL of the chatbot application within the webpage's CSS/programming code reliably identifies the chatbot vendor.
  • Classified as probably not a substantial claim.

    Study design and conductreasoning

    The study analyzed only the home page or landing page of each institution.

    “Accordingly, we chose to only analyze the home page or landing page.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Community colleges receive less public funding than four-year institutions.

    “community colleges receive less public funding than four-year institutions (Kelchen, 2019)”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kelchen, 2019 source
  • May combine two claims.

    Premisecitation

    Many community college students, including Latinx students, work full-time and enroll part-time because of their multiple responsibilities, and thus may not have time during the day to ask questions.

    “Research has also shown that many community college students, including Latinx students, work full-time and enroll part-time because of their multiple responsibilities, thus they may not have time during the day to ask questions (American Association of Community Colleges [AACC], 2022; Alcantar & Hernandez, 2020).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites American Association of Community Colleges [AACC], 2022

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Alcantar & Hernandez, 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

    Premisecitation

    HSI community college leaders have the potential to set the standard of how to best integrate servingness for linguistically diverse students and their families.

    “HSI community college leaders have the potential to set the standard of how to best integrate servingness for linguistically diverse students and their families (Garcia, 2019).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Garcia, 2019 source
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationown data

    The interaction with live agents humanizes the experience for Latinx students and families who are trying to learn more about an institution.

    “The interaction with live agents humanizes the experience for Latinx students and families who are trying to learn more about an institution.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Garcia and Ramirez (2018) encourage HSI leaders to become empowerment agents who are critically conscious of how they are serving students.

    “Garcia and Ramirez (2018) encourage HSI leaders to become empowerment agents who are critically conscious of how they are serving students.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Garcia and Ramirez, 2018

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Providing Spanish and bilingual resources and services is one way to enhance servingness.

    “Providing Spanish and bilingual resources and services is one way to enhance servingness”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The study suggests that HSI community colleges can use AI technology to enhance their HSI designation to better serve Latinx students.

    “Our study suggests that HSI community colleges can use AI technology to enhance their HSI designation to better serve Latinx students”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    By investing in bilingual staff and online resources, HSI community colleges can implement servingness through their websites.

    “By investing in bilingual staff and online resources, HSI community colleges can implement servingness through their websites.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Premisecitation

    One part of servingness is creating validating spaces and interactions for Latinx students, which must also include online and virtual platforms.

    “One part of servingness is creating validating spaces and interactions for Latinx students (Garcia et al., 2019), which must also include online and virtual platforms.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Garcia et al., 2019 source
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    HSI servingness would require that HSI community colleges create validating and inclusive virtual spaces not only for those who prefer to communicate in Spanish or English but also those who prefer to communicate in other languages.

    “More importantly, HSI servingness would require that HSI community colleges create validating and inclusive virtual spaces not only for those who prefer to communicate in Spanish or English but also those who prefer to communicate in other languages.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    Providing the option to communicate in Spanish shows Latinx students that the institution values their culture and background.

    “Additionally, providing the option to communicate in Spanish shows Latinx students that the institution values their culture and background.”
    Quote in the paperNot a substantial claim
34 of 34
  • The paper relies on this without support.

    The host URL of the chatbot application within the webpage's CSS/programming code reliably identifies the chatbot vendor.

    “Finally, we used the inspect element feature to explore the cascading style sheets (CSS) coding of the webpage to determine whether a particular chatbot company was responsible for writing the chatbot script, revealed through the host URL of the chatbot application within the programming code.”
  • The authors state this as an assumption.

    Ocelot would be the most user-friendly chatbot because of its high cost.

    “Because of the high cost, we assumed Ocelot would be the most user-friendly chatbot”
  • The paper relies on this without support.

    Some Latinx students and their families may not feel comfortable sharing personal information, particularly if they are undocumented.

    “some Latinx students and their families may not feel comfortable sharing this information, particularly if they are undocumented.”
  • The paper relies on this without support.

    Students may not be fully aware who has access to their personal data collected by chatbots.

    “students may not be fully aware who has access to this data (e.g., other higher education institutions, chatbot companies).”
  • The paper relies on this without support.

    Undocumented students and families may be hesitant to share personal data that may make them vulnerable to federal immigration officials.

    “This is particularly a concern for undocumented students and families who may be hesitant to share personal data that may make them vulnerable to federal immigration officials.”
5 of 5

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • The study population is reported as both 234 and 230 colleges

    Minor

    The abstract says the study examined all 234 HSI community colleges on the HACU list, but the methods identify and analyze 230. This changes the population underlying the reported prevalence of chatbots.

1 of 1

Cites

Flagged first, then by citations

Research works printed but not linked: 2

  1. Banerjee B. (2021). Smart student guide case of HUGI a chatbot for education. (Publication No. 10024) [Master’s thesis Haaga-Helia University of Applied Sciences]. RefWorks. http://urn.fi/URN:NBN:fi:amk-2021060915260
  2. Sharma R.M. (2019). Chatbot based college information system. International Journal of Multidisciplinary 4(3) 109–112. https://rrjournals.com/wp-content/uploads/2019/03/109-112_RRIJM190403024.pdf

Cited by

Sorted by citations

Authors
Hernández, S. H., Taylor, Z. W., & Ortega, G.
Affiliations
  • Northern Arizona University (US)
  • University of Southern Mississippi (US)
  • Texas Tech University (US)
Corresponding author
Hernández, Susana H.
Venue
Community College Review
Published
2025-05-07
Volume
53
Issue
4
Pages
434-452
ISSN
  • 0091-5521 (ISSN-L)
  • 1940-2325
OpenAlex
W4410186605
Cited by
1
References
46
Citations by year
2026: 1
FWCI
0.10
Type
journal article
Language
en
Open access
closed
Locations
Community College Review (publishedVersion)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › AI in Service Interactions
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
  • Social Sciences › Education › Higher Education Research Studies
Keywords
AI chatbots, Hispanic-serving institutions, community colleges, servingness, language equity, bilingual services, chatbot usability, institutional websites

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-11.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-11.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review