<p>Generative artificial intelligence (GenAI) has catalyzed renewed institutional attention to academic integrity, often framed as a crisis of cheating and surveillance. Yet this framing obscures how GenAI can function as a compensatory tool for executive-function demands that higher education routinely treats as invisible prerequisites for “independent” academic work. This matters acutely for students with untreated or under-treated attention-deficit/hyperactivity disorder (ADHD), whose access to diagnosis, medication, and coaching remains uneven across gender, race, socioeconomic status, and geography. Drawing on a critical integrative review and disability-studies-informed policy analysis, we synthesize evidence on (a) contemporary academic integrity responses to GenAI, (b) ableism embedded in assessment cultures and accommodation regimes, (c) barriers shaping ADHD under-treatment, and (d) emerging findings on GenAI use by disabled students. Results are organized into five themes: integrity discourse as compliance; GenAI as executive-function scaffolding; accommodation systems as ableist gatekeeping; detection and surveillance as high-risk enforcement; and an emergent shift toward authorized assistance and transparency frameworks. We propose an Access-Centered Academic Integrity (ACAI) model that distinguishes supportive from substitutive GenAI use, emphasizes assessment validity over detection, and implements universal options that reduce pressure to disclose disability. We conclude with a policy toolkit for course and institutional levels, including authorized-use matrices, disclosure gradients, and assessment redesign principles that foreground equity, accessibility, and pedagogical integrity.</p>