Consortium for School Networking (CoSN)

The Office for Civil Rights (OCR) in the U.S. Department of Education has provided guidance on how AI use in schools can create or contribute to discrimination. They lay out illustrative scenarios of how AI use might violate federal civil rights laws. This is not an excuse to bury our heads in the sand out of fear, but an opportunity to engage in conversations so all citizens can develop their own decision-making framework for AI's proper use (aka AI Literacy). Here are some of the scenarios. They are all too real. ? A teacher uses an AI-detection tool to evaluate student essays, which incorrectly flags the work of English Learner (EL) students as AI-generated, leading to failing grades and academic dishonesty charges despite their innocence. When a parent appeals, the principal defends the tool’s accuracy, raising concerns about discrimination against EL students and prompting a potential OCR investigation for unequal participation in the instructional program. ? A high school student discovers a social media account posting AI-generated deepfake nudes of classmates, sparking widespread discussion at school. Although the school is aware of the account and reports it to the police, it takes no further action to address the harassment, raising concerns of prohibited harassment and inadequate response that could prompt an OCR investigation. ? A school district implements facial recognition technology that inaccurately flags Black students as persons of interest, leading to repeated questioning by school staff and missed class time. Despite student complaints and resulting peer harassment, the school fails to address these issues, creating a hostile environment and raising concerns about racial discrimination that could... you guessed it, prompt an OCR investigation. My recommendation is to read the scenarios and examine not only whether AI was being used properly based on its capabilities and limitations but also whether the humans involved made the right decisions based on the AI output and recommendations. Legal repercussions are not the only reason to take the examples in this document seriously. People can use this document?and the recently released Office of Ed Tech toolkit to build their AI literacy so that they can avoid and mitigate risks while also seeking beneficial ways of using AI. #TeachAI to avoid discriminatory use cases