Skip to content
Apart Research
Apart Research
Contact us
Sprint projectJun 21, 2026Harare, Zimbabwe

AfroJailbreak-ZW: Evaluating Jailbreak Resistance in Shona

Malvin T. Machingura · Team LoneStar

Submitted to Global South AI Safety Hackathon. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: AfroJailbreak-ZW: Evaluating Jailbreak Resistance in Shona

Open in a new tab (opens in new tab)
Share

A pilot study testing whether ChatGPT and Gemini are easier to jailbreak in Shona than in English. In a small test (n=5–6 per prompt type), both models complied with harmful requests more often in Shona and Shona-English code-switched prompts than in English, suggesting AI safety protections built mainly for English may not hold for Zimbabwean languages

Reviews

Judging this Sprint?

Review this project

Your public critique appears on this page without your name. Your private critique is not published; only the Apart team reads it. If you agree below, we share your review with grantmaking.ai (opens in new tab) and the Transformative AI Fund so strong projects can be funded.

Not shown on this page.

Shown on this page, without your name.

Only the Apart team reads this, and funders if you agree below.

Share my name publicly on grantmaking.ai *
Share my private critique with funders *

How much would this matter for AI safety if it worked? How innovative is it? For scores of 4-5: is this actually new to the field, or replicating recent work?

Scoring guide
  1. 1Negligible. No clear problem addressed, or no meaningful novelty.
  2. 2Limited. Addresses a real problem but with a generic or well-trodden approach. Incremental at best.
  3. 3Moderate. Clear problem with a reasonable approach; some novelty in framing or method beyond routine application of existing tools.
  4. 4Significant. Important problem with an original approach, or identifies a neglected problem area. A valuable contribution others could build on.
  5. 5Exceptional. Tackles a critical AI safety problem with a genuinely novel approach, or opens a new research direction. Clear theory of change. You'd be excited to share this with researchers in the area.

How sound are methodology, implementation, and findings?

Scoring guide
  1. 1Seriously flawed. Methodology broken, results uninterpretable, or implementation doesn't work.
  2. 2Weak. Approach has significant gaps: missing validation, flawed experimental design, or incomplete implementation.
  3. 3Competent. Technically solid given the short duration. Methodology makes sense, results are interpretable, limitations acknowledged, work builds toward clear conclusions.
  4. 4Strong. Thorough methodology with convincing validation. Results clearly support conclusions. Immediately useful for future work.
  5. 5Exceptional. Ambitious scope executed rigorously. Surprising findings, novel methods, or unusually robust validation.

How clearly are work, findings, and impact potential communicated?

Scoring guide
  1. 1Incomprehensible. Cannot determine what the project is actually claiming or doing.
  2. 2Hard to follow. Key information buried, missing, or diluted by excessive length. Significant effort to extract main points.
  3. 3Clear enough. Can understand the problem, approach, and results without undue effort. Core content clearly present: problem, method, findings, limitations.
  4. 4Well presented. Easy to follow, well-structured, appropriate level of detail. Target audience would get it quickly.
  5. 5Exceptionally clear. A pleasure to read. Complex ideas made accessible. Could serve as a model for how to present this type of work.

  1. The project has chosen a good idea of identifying and resolving jailbreak in low-resource languages. It introduces an additional metric LVI aiming to make ASR more accurate by considering the relevance and comprehension factor. However, the LLMs in test(ChatGPT/Gemini) don't claim to support Shona. So, the baseline of testing isnt accurate. Also, the methodology uses only 6 queries for testing, which is very small and can highly misguide the accuracy/orientation of the whole project.

  2. This is a valuable and timely pilot, offering the first jailbreak‑resistance evaluation for Shona and introducing a Language Vulnerability Index that adjusts attack success rates for relevance/comprehension, which is highly pertinent for Global South safety where low‑resource languages and code‑switching are common.

    The methodology—three parallel prompt sets (English, Shona, code‑switched), multiple attack categories, GPT vs Gemini comparison, and an open harness that can be extended to other Southern African languages—is well thought out for hackathon scale, but quantitative strength is limited by very small sample sizes (n=5–6), single‑rater labelling, and lack of independent back‑translation or inter‑rater reliability.

    The paper is clearly written, with honest discussion of limitations and dual‑use risks, and would be strengthened by slightly more detail on how LVI is computed, a clearer plan for scaling the dataset and judge pool, and an explicit roadmap for operationalising this benchmark with local institutions and providers.

    Read full reviewShow less

  3. 3 / 2 / 3

    Useful first step for Shona AI safety. The main issue is that the sample size is very small, so the results should be treated as early evidence.

Cite this project

@misc{machingura2026afrojailbreakzw,
  title = {{AfroJailbreak-ZW: Evaluating Jailbreak Resistance in Shona}},
  author = {Malvin T. Machingura},
  year = {2026},
  month = jun,
  note = {Submitted to Global South AI Safety Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/afrojailbreakzw-evaluating-jailbreak-resistance-in-shona-7bmu}},
  url = {https://apartresearch.com/sprints/projects/afrojailbreakzw-evaluating-jailbreak-resistance-in-shona-7bmu}
}

More from Global South AI Safety Hackathon

All 216 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026