Response to Ariel and Hasid’s Letter

We would like to thank Drs. Ariel and Hasid for their thoughtful review of our article and their engagement on the topic of large language models (LLMs) for support for patients with alcohol misuse. Their recent letter raises several important points. For example, future work should delineate the potential clinical severity of problematic advice and assess performance in additional languages. Our raters catalogued examples of incorrect information, but we did not attempt to quantify risk. This was in part because risk varies by patient factors (e.g., co-occurring conditions, medications) to which LLMs typically lack access.