Lately, there have been so many teenage pregnancies across Kenya, wra kenya advocates for listening to women and girl voices for their lived experiences in srhr through asking listen, and act model. How should they approach this issue through their model of operation

GPT-5.4

Answer: I maintain my position: WRA Kenya should use its Ask, Listen, and Act model to center girls’ lived experiences, convert those insights into locally tailored SRHR services and protections, and use the resulting evidence to hold families, schools, health systems, and government accountable.

The other models largely reinforced rather than changed my view. They converged on the same core principle I already emphasized: teenage pregnancy in Kenya should be approached as a structural SRHR issue, not merely an individual moral failing, and the model works best when girls’ voices directly shape services, advocacy, and accountability. What I found most useful in the other answers were refinements in emphasis, especially Claude Opus 4.7’s stronger articulation of system accountability and its concrete examples of where pressure should be applied: county health teams, schools, local leaders, and implementation gaps in adolescent SRHR policy. That sharpened the governance dimension, but it did not fundamentally alter my position because my original answer already argued that WRA Kenya should move from lived experience to institutional accountability across families, schools, health systems, and government. Grok 4.20 and Gemini 3.1 Pro also strengthened the case for youth-led, stigma-free engagement and co-designed responses, but again these points are consistent with my prior answer rather than persuasive grounds for changing it. I already argued for confidential, age-appropriate, locally adapted engagement and for translating what is heard into tailored interventions and protections. So I would keep my position, while noting a few useful additions from the roundtable: WRA Kenya should ensure that girls who are already pregnant or parenting are included, not sidelined; it should institutionalize feedback loops so participants see what changed; and it should push not only for services like contraception and youth-friendly care, but also for school re-entry, protection from coercion and violence, and county-level budget and policy accountability. In short, the best approach remains the same: Ask widely and safely, Listen seriously and analytically, and Act in ways that are co-designed with girls and aimed at both immediate support and systemic reform.

Gemini 3.1 Pro (changed mind)

Answer: WRA Kenya should apply the Ask, Listen, and Act model by co-creating trauma-informed, safe spaces to gather girls' lived experiences, ensuring robust safeguarding and referral pathways, and translating these insights into systemic accountability and youth-tailored SRHR policies.

I have updated my perspective based on GPT-5.4's critical insight regarding the necessity of trauma-informed practices and safeguarding when engaging with vulnerable adolescents. Gathering stories about teenage pregnancy inevitably surfaces traumatic experiences, such as sexual violence, coercion, or abuse, meaning that extraction of stories without protection can do more harm than good. During the 'Ask' phase, WRA Kenya must move beyond merely providing non-judgmental spaces. They must establish rigorous safeguarding protocols and clear referral pathways for legal aid, psychosocial support, child protection, and emergency healthcare. Questions should be co-designed with adolescents to sensitively navigate complex realities like transactional sex, poverty, or lack of consent, ensuring the process is driven by the girls' comfort and safety. In the 'Listen' phase, active listening must be analytical, serious, and intersectional. By systematically documenting these lived experiences and disaggregating the data across different demographics—such as rural communities, urban informal settlements, or girls with disabilities—WRA Kenya can pinpoint the specific, localized barriers to SRHR. This treats the girls' voices as credible, undeniable evidence rather than mere anecdotes. Finally, the 'Act' phase must leverage this evidence to demand concrete accountability across families, health systems, and government. WRA Kenya should partner directly with the girls to advocate for actionable reforms, such as the enforcement of school re-entry policies for teen mothers, consistent funding for youth-friendly contraceptive services, and community-wide campaigns to dismantle stigma. Crucially, closing the feedback loop by showing the girls how their voices led to tangible change is essential to validate their courage and prove the model's efficacy.

Grok 4.20

Answer: WRA Kenya should deploy their Ask, Listen, and Act model by creating safe, youth-led spaces for teenage girls to share experiences, systematically analyzing these with broader stakeholder input to identify root causes, and translating them into community-driven SRHR programs, youth-friendly services, and policy reforms.

My original position emphasized gathering authentic testimonies from girls, translating them into programs addressing poverty, education gaps, and contraceptive access, and maintaining feedback loops for relevance. This remains the most balanced application of the model. Reviewing the other responses shows strong alignment on core principles: all stress safe, confidential, non-judgmental spaces, avoiding stigma, co-creation with adolescents, and multi-level action spanning services, schools, communities, and government accountability. GPT-5.4 adds valuable detail on trauma-informed practices, safeguarding, disaggregated data, and outcome measurement that would strengthen implementation without altering the fundamental approach. Gemini 3.1 Pro usefully highlights grassroots mobilization and peer dialogues, while Claude Opus 4.7's Listen-Ask-Act variation and focus on interrogating failing systems provide a helpful structural lens. These contributions enrich the execution steps but do not convince me to change the headline stance or reorder the model to begin with Listen, as the organization's stated framework centers on 'asking' first to elicit voices directly. The integrated approach avoids tokenism by keeping girls at the center as both informants and co-designers, ensuring actions target interconnected drivers like coercion, service stockouts, and harmful norms. Continuous feedback loops with participants will prevent the process from becoming extractive, while measurable indicators around service uptake, school retention, and attitude shifts will demonstrate real impact. This maintains fidelity to the Ask, Listen, and Act model while incorporating the best insights from the roundtable.

Claude Opus 4.7 (changed mind)

Answer: WRA Kenya should apply its Listen–Ask–Act model by first creating safe, youth-led spaces to listen to pregnant and at-risk girls, then asking structural questions about why SRHR systems are failing them, and finally acting through community-anchored advocacy that pairs girls' testimonies with policy and service reforms — but this must be reinforced with trauma-informed safeguarding, co-creation with adolescents, and a measurable accountability loop.

I maintain my core position because the three-stage logic I outlined — targeted listening in high-burden counties, structural interrogation of duty-bearers, and co-designed action — remains the most practical application of WRA Kenya's model. However, GPT-5.4's response convinced me to strengthen two dimensions I underweighted in round 1, and I want to integrate them explicitly rather than pretend my original answer was complete. First, trauma-informed safeguarding. My original 'Listen' phase emphasized safe spaces and young facilitators, but GPT-5.4 rightly pointed out that girls will disclose rape, incest, transactional sex, and abuse by authority figures during these sessions. Without pre-established referral pathways for psychosocial support, legal aid, child protection, and emergency post-rape care (PEP, EC, PRC forms), the listening process itself can retraumatize or endanger participants. WRA Kenya must therefore build safeguarding infrastructure before the first circle convenes — not as an afterthought. Second, closing the accountability loop with measurable indicators. My 'Act' phase focused on advocacy outputs (testimonies plus policy asks), but GPT-5.4's framing of Ask–Listen–Act as a 'rights-based accountability cycle' with indicators like youth-friendly service uptake, school re-entry rates, respectful-care experience, and feedback to source communities is sharper than mine. Without this, WRA Kenya risks extractive listening — taking girls' stories upward without returning results downward. I did not adopt Grok 4.20's or Gemini 3.1 Pro's reordering of the phases because the user explicitly named the model as 'listen, and act' with 'ask' as the diagnostic middle step, which matches my original sequencing. I also maintain my emphasis on centering already-pregnant and parenting girls — the most silenced group — which none of the other models foregrounded as strongly. Boys and men remain secondary listeners, not co-primary voices, because the question is about centering girls' lived experiences. So: same model, same sequence, same county targeting — but now with explicit safeguarding pre-conditions and a measurable feedback loop, thanks to GPT-5.4.

Final Summary

The AI Roundtable recommends centering the "Ask, Listen, and Act" model on safe, youth-led spaces where girls share lived experiences to drive systemic policy reform. GPT-5.4 and Claude Opus 4.7 stressed that this process must be trauma-informed, providing immediate legal and medical support to those disclosing abuse. Gemini 3.1 Pro and Grok 4.20 emphasized that these narratives should be used to co-design youth-friendly health services and hold local governments accountable.

All 4 models agreed