Artificial intelligence technology will replace human evaluators on this year’s State of Texas Assessments of Academic Readiness exams, saving state authorities over $15 million.
The Texas Education Agency (TEA) will use automated scoring engines to grade students’ written answers on the 2023-2024 STAAR exams, a series of standardized tests by grade level and subject used to determine — among other things — district performance, reported The Texas Tribune. The computers employ a natural language processing technology similar to that seen in AI chatbots. Several safeguards will be in place to ensure that the technology is grading answers correctly.
“We have always had very robust quality control processes with humans,” said Chris Rozunick, TEA’s division director for assessment development, per The Tribune.
After computers have gone through and scored all the written answers, a quarter of them will be reassessed by human graders to ensure quality. Computers can also flag scores as “low confidence” if they encounter something unique during their assessment, automatically assigning the responses to human graders.
The shift toward what TEA refers to as “hybrid scoring” was announced last December, and it has raised concerns, as reported in The Dallas Express.
For instance, Rep. Gina Hinojosa (D-Austin), a House Public Education Committee member, referred to the news as “surprising” in a letter addressed to TEA Commissioner Mike Morath.
“I do not recall ever receiving notice of this novel and experimental method for grading high-stakes STAAR tests,” she wrote, according to The Dallas Morning News.
TEA’s accountability system grades school districts and individual campuses on an A-F scale depending on several metrics, including students’ STAAR exam results. Chronic underperformance can lead to state intervention, as exemplified by the state takeover of the Houston Independent School District.
Last year, TEA adjusted this accountability system, resulting in several Texas school districts filing a lawsuit, keeping their latest accountability data under wraps. Dallas ISD — which has seen considerable academic struggles in the past few years — is among the plaintiffs, with its accountability report from the 2021-2022 school term being the latest available.
TEA also redesigned the STAAR exam to mimic classroom conditions and assess learning more closely, which has resulted in the shift to considerably more open-ended questions.
“We wanted to keep as many constructed open-ended responses as we can, but they take an incredible amount of time to score,” said Jose Rios, TEA director of student assessment, per The Tribune.
Thanks to computer scoring, Rios reported that TEA would need to hire no more than 2,000 temporary scorers — a significant drop from the 6,000 brought on through a third-party contractor in 2023.
Yet educators still have concerns that some students will be penalized by hybrid scoring — for instance, those writing with more originality or non-native English speakers — or that it will increase anxiety among students and teachers alike.
“There’s always this sort of feeling that everything happens to students and to schools and to teachers and not for them or with them,” remarked Carrie Griffith, policy specialist for the Texas State Teachers Association, per The Tribune.