Perplexity AI’s CEO publicly warned students not to use the company’s Comet browser to complete coursework after a viral video showed the agent finishing a Coursera assignment in seconds. The episode exposed academic-integrity vulnerabilities as agentic browsers can autonomously navigate websites, fill forms and execute tasks—raising concerns about remote assessment, plagiarism detection and vendor risk. At the same time, researchers and enterprise leaders warn of automation bias and the need to centralize oversight: as institutions deploy AI agents, human over-reliance and accountability gaps can produce errors or ethical lapses. Campus IT and academic affairs offices face dual tasks—implementing guardrails and educating students and faculty about appropriate AI use. Universities must update honor codes, assessment design and IT procurement processes to address agentic tools that can both aid learning and enable misuse. Legal, privacy and pedagogical teams should coordinate on detection, remediation and student support.