The rise of AI generated texts offers promise but creates new challenges for social work teaching. A recent survey found that 89% of higher education students used AI on their homework. AI generated text may be difficult to distinguish from a student's own work, yet are being submitted as the student's own work. This poses new challenges to definitions of plagiarism in education and in professional publication, as well for policies addressing this issue. Identifying AI generated text is difficult and requires critical thinking. While some software options claim effectiveness, recent tests show their results are mixed. Recommendations are offered for assignment development and policy changes to address this new concern in social work education and in publications.