Suicide is a significant societal issue; despite substantial scientific and clinical efforts in recent decades, the assessment of suicide risk still falls short of providing sufficiently predictive results and also faces significant challenges in accessibility to the general public. This presentation will describe the capabilities of generative artificial intelligence, specifically large open-access language models, for assessing suicidal thoughts and behaviors in comparison to mental health professionals. It will highlight encouraging preliminary findings on the ability of advanced models, such as CHATGPT-4, to perform assessments that closely resemble those of professionals. Additionally, the presentation will explore attempts to understand if and how large language models incorporate various risk factors (such as previous suicidal attempts, access to weapons, depression, etc.) to carry out suicide risk assessments. Furthermore, results will be presented that demonstrate the application of these capabilities, among other uses, to foster responsible media coverage of suicide issues with the assistance of large language models. Lastly, the presentation will also describe several methodological approaches to research on the topic of suicide with generative artificial intelligence and will address ethical questions in research and clinical practice involving the integration of large language models.