Abstract:

A shift in demographics in the United States of America (USA), at a time where English is now the national language, combined with a decline in literacy scores, is increasing for educators. A 2025 National Assessment of Educational Progress shows the average score for grade 12 students is an historically documented low: 32% of high school seniors score below the basic reading proficiency level. With the ubiquity of Generative Artificial Intelligence (GAI), students are availing of the ease of use and rapid response time of AI to supplement or even replace basic academic skills for reading or writing.

The birth rate in the USA is rising in the Hispanic/Latino community and decreasing in the White community. The home language; the language spoken in student homes, for many native USA MLLs is predominantly Spanish. Those learners are potentially immediately disadvantaged when entering the public education setting. In 2019, one in five families spoke a language other than English in their homes, a 194% shift over 45 years.

High school seniors are entering higher education and are unprepared for higher education. Faculty are tasked with examining student work products, trying to ascertain if submissions reflect student ability, or if the submissions are AI supported. In 2024, MLLs said that "ChatGPT did a better job than we could have done”, and that “Professors expect our English to be perfect”. There has historically been a perspective that MLLs in higher education disadvantage native English speakers. However, recent research indicates faculty concerns about the quality of work from students of all backgrounds, and the increasing reliance on AI without student comprehension of output, nor necessary literacy skills to accurately provide and evaluate GAI output. These concerns are grounded in the decline of literacy at a national level, with disparity between states. For international students in higher education, challenges arise with academic and/or cultural norms in the USA, for academic success: internal, academic, and family/cultural.

This paper builds on previous work addressing Multiple Language Learners (MLLs) in higher education, and the opening of the AI Pandora’s box. For the scope of this research MLLs defined as students who have attended public or private schools in the USA, for whom English is not their first language. Through qualitative interviews, MLLs' data is gathered on AI usage. Participants were provided two tasks; one independent, and one with AI support. The outcome in terms of time completion was markedly higher when done independently, and the task score was significantly lower.

Future research, given the literacy level of native English speakers, should focus on the replication of this research in the native English speaker student population to assess the validity of faculty concern that independent writing and comprehension skills are declining, and students are using AI to succeed. Without strong fundamental English speaking and writing skills, the ability to create meaningful input, research and validate the output recognizing bias and inaccuracies is minimalized. The onus will be on educators to teach students literacy skills and how those skills are applied to Generative AI.

Keywords:

Generative AI, Literacy, Assessment, Multiple Language Learners, Education.