Skip to content
Apart Research
Apart Research
Contact us
Sprint projectJun 21, 2026lusaka, Zambia

Ai fraud zambia research paper

JOHN KAMFWA · Team Young Tecsperts

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

Read the report

Report: Ai fraud zambia research paper

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

This research investigates how AI-powered tools (deepfakes, face synthesis, image generation) are enabling large-scale fraud targeting Zambian citizens. Through interviews with victims, law enforcement, and civil society, analysis of reported fraud cases, and development of a detection framework, we identify how scammers use AI-generated profiles for romance scams, fake product images for e-commerce fraud, and deepfaked content for impersonation. We develop a lightweight detection tool optimized for low-bandwidth, offline mobile use and achieve 80% accuracy on real Zambian fraud cases. Key findings: victims miss behavioral and textual red flags when emotionally vulnerable; existing Western-trained detection models perform poorly on compressed Zambian mobile images; user education significantly improves fraud detection. We recommend technical tools, user education, platform verification mechanisms, and institutional cybercrime capacity-building as coordinated solutions.

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. Good idea, approach could have been to solve the identity verification for AI instead of engineering on the fraud side. I like how you think about the problem in depth but starting at one point and providing one solution is important.

  2. Strengths: Addresses a practical real-world problem with significant societal value. The project demonstrates a clear understanding of the domain and presents a feasible solution. Areas for Improvement: Provide more evidence of model performance through precision, recall, false positive rates, or other relevant evaluation metrics. Discuss data quality, bias mitigation, scalability, and deployment challenges to better demonstrate production readiness.

  3. The project is well-executed and addresses an important problem. However, there are two notable limitations:

    - The methodology section describes an ambitious mixed-methods design (fraud case dataset from police and NGOs, interviews with victims and moderators, user testing with 20/20 synthetic-vs-authentic profiles, and an ensemble CNN + text + behavioral detector), but the report never states the sample sizes actually achieved — how many cases were collected, how many interviews conducted, how many user-test participants recruited. The reported detection metrics (~88% precision on public datasets, ~80% on Zambian cases) are presented without confusion matrices, test-set composition, or baseline comparisons, making it difficult to distinguish what was empirically measured from what is proposed.

    - The scope is very broad — romance scams, e-commerce fraud, impersonation, deepfakes, user education, platform policy, and legal recommendations — but no single strand is developed with depth sufficient to produce a defensible finding. The technical detection framework, the qualitative victim interviews, and the policy recommendations each warrant separate treatment; combined into one report, each is reduced to a summary, and the resulting recommendations (be skeptical, demand video calls, implement seller verification) are generic enough that they could have been written without the described empirical work.

    Read full reviewShow less

Cite this project

@misc{kamfwa2026ai,
  title = {{Ai fraud zambia research paper}},
  author = {JOHN KAMFWA},
  year = {2026},
  month = jun,
  note = {Submitted to Global South AI Safety Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/ai-fraud-zambia-research-paper-3yfr}},
  url = {https://apartresearch.com/sprints/projects/ai-fraud-zambia-research-paper-3yfr}
}

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