Young people's perceptions and recommendations for conversational generative artificial intelligence in youth mental health

Referee: [Methods] Methods section: The process for developing the four themes from workshop data is not described, including the thematic analysis approach, coding procedures, inter-rater reliability checks, quote selection criteria, or saturation assessment. This directly weakens the central claim that the themes accurately reflect participant perspectives and can inform system requirements.

Authors: We acknowledge this omission and agree that greater detail is required. In the revised manuscript, we will expand the Methods section to fully describe our reflexive thematic analysis process following Braun and Clarke (2006). This will include the six-phase approach, how initial codes were generated inductively from the workshop transcripts by two researchers, the collaborative process for theme development and refinement (including regular team discussions to resolve differences in interpretation), criteria for selecting representative quotes (ensuring they illustrated the theme while capturing a range of participant voices), and our approach to thematic saturation (iterative analysis until no new themes or sub-themes emerged). We will clarify that formal inter-rater reliability statistics are not standard practice in reflexive thematic analysis, which prioritizes interpretive depth; instead, we relied on a transparent, team-based consensus process. These additions will better substantiate how the themes were derived from the data and support their use in informing system requirements. revision: yes

Referee: [Results] Results and participant description: No demographics, recruitment criteria, geographic/cultural diversity details, or inclusion/exclusion criteria are reported for the 32 participants. This undermines the abstract's and discussion's extension of findings to 'young people's perceptions' and 'broader youth needs' in youth mental health.

Authors: We agree that this information is essential for contextualizing the findings. We will add a dedicated 'Participants' subsection detailing the recruitment strategy (through Australian youth mental health services and targeted online channels), inclusion criteria (young people aged 16–25 with an interest in mental health and technology), exclusion criteria (e.g., acute distress preventing informed consent), and available demographic characteristics (age range, gender identity distribution, and cultural background where reported). Due to ethical constraints and privacy protections, granular individual-level data were not collected. We will also revise the abstract and discussion to explicitly frame the results as insights from this specific Australian sample rather than broad claims about 'young people's perceptions' or 'broader youth needs,' thereby avoiding overgeneralization. revision: yes

Referee: [Discussion] Discussion: The translation of co-design outputs into 'key implications for service integration' and recommendations for ethics/design/governance lacks any validation steps, pilot testing, or explicit discussion of generalizability limits, making the actionable claims load-bearing but unsupported.

Authors: We accept this critique and will strengthen the Discussion accordingly. We will add an explicit limitations subsection that addresses generalizability, highlighting the modest sample size, the Australia-specific setting, and the exploratory nature of the co-design workshops. We will clarify that the implications for service integration, ethics, design, and governance are presented as direct outputs from the young people's co-design recommendations and are intended to inform future work rather than as validated or immediately implementable guidelines. We will note that pilot testing and further validation fall outside the scope of this qualitative study but represent important next steps. These changes will ensure the claims are appropriately supported and caveated while preserving the value of the youth-centered insights. revision: yes