Texas AI Grading System
Texas students taking their state-mandated exams are now being graded by an artificial intelligence-powered scoring system, marking a significant shift away from human graders. The Texas Education Agency (TEA) is implementing an “automated scoring engine” that utilizes natural language processing to grade open-ended questions on the STAAR exams, aiming to save $15–20 million annually by reducing the need for temporary human scorers. However, concerns have been raised by educators about potential inaccuracies, as evidenced by a previous trial that saw a “drastic increase” in zero scores for constructed responses, prompting questions about the reliability of the new system. Texas AI Grading System.
TEA’s Director of Student Assessment, Jose Rios, told the Tribune the state hired around 2,000 human scorers this year, a small margin compared to the 6,000 employed in 2023.
According to a scoring report by the TEA, students’ responses will be graded first by a computer. A hired human scorer will grade roughly 25 percent of those responses. If the computer has a “low confidence” score, it will also be re-scored by a human. Some tests are also up for review if the computer’s programming catches unrecognizable responses like slang words, phrases, or languages other than English. “The purpose of this routine is to ensure that unusual or borderline responses receive fair and accurate scores,” TEA wrote in the December report.
Request a Human Rescore
According to the Tribune, students and parents who disagree with the computer and human scores can request a rescore for $50. The fee can be waived if the computer’s or human’s score is wrong. TEA guidelines state that the STAAR test measures how much a student has learned about a subject. STAAR test scores also largely determine how the state’s education agency grades districts and campuses through its A-F accountability system. Students in Texas start STAAR testing beginning in 3rd grade.
While TEA emphasizes the differences between its closed system and general artificial intelligence, the move underscores broader debates around the use of AI in education and its potential impact on students. As AI-powered tools become more prevalent, educators and students alike grapple with questions of fairness, accuracy, and transparency, particularly in high-stakes assessments like standardized tests. The implementation of AI grading systems in education reflects both the promise and pitfalls of technology in shaping the future of learning and evaluation.
Source: Texas AI Grading System