Hikayat - Interactive Stories to Learn AI Safety
Chaitanya Mittal, Masah Arar, Nermeen Rizwan, Saima Tariq Khan · Team Dune Tech
Submitted to Women in AI Safety Hackathon. Sprint projects are early-stage work by participants, not Apart Research publications.
This paper presents an interactive, scenario-based learning approach to raise public awareness of AI risks and promote responsible AI development. By leveraging Hikayat, traditional Arab storytelling, the project engages non-technical audiences, emphasizing the ethical and societal implications of AI, such as privacy, fraud, deepfakes, and existential threats. The platform, built with React and a flexible Markdown content system, features multi-path narratives, decision tracking, and resource libraries to foster critical thinking and ethical decision-making. User feedback indicates positive engagement, with improved AI literacy and ethical awareness. Future work aims to expand scenarios, enhance accessibility, and integrate real-world tools, further supporting AI governance and responsible development.
Reviews
I love that you've taken inspiration from your culture towards AI safety education!
Here are some things I particularly enjoyed:
- I liked your approach of picking out particular themes in AI safety you wanted to convey and building a story around it.
Here are something I thought could have been improved:
- I would have loved for the story to be more fleshed out, but I understand the lack of time! In the future, I think it would be better to invest time into the actual story which is your unique proposition and do a simple technical mockup, perhaps even just using Google Slides, so that you could invest more time in building out these impactful stories.
- I appreciated that you got feedback directly from users! Given the few respondents you had, it might have been best to conduct user interviews with them instead of run a quantitative survey. I would recommend using the Mom Test to guide that.
I'd love to see these stories more fleshed out to help broader audiences understand the impact of AI!a
Read full reviewShow less
What a cool project! The problem statement was well developed and the potential of Hikayat to solve a need relating to public education was clearly articulated. Some references felt imprecise (e.g. it's unclear from the aisafety.info link provided where the selected themes were drawn from). In some parts of the paper, it was hard to understanding what was drawn from a reference and whether further content was available in the appendices (e.g. elaborating on the findings from the review of existing AI safety curricula in an appendix may be appropriate, to understand your process and which themes you drew from the analysis). The scenarios themselves are compelling, and with further user-testing and additional samples, I can imagine engaged storytelling and discussion across a variety of formats. Consider adding research papers to the further reading section, outlining the research on current capabilities and limitations, or theoretical arguments relating the events in the stories. I'm excited to see where this goes.
Read full reviewShow less
Cite this project
@misc{mittal2025hikayat,
title = {{Hikayat - Interactive Stories to Learn AI Safety}},
author = {Chaitanya Mittal and Masah Arar and Nermeen Rizwan and Saima Tariq Khan},
year = {2025},
month = mar,
note = {Submitted to Women in AI Safety Hackathon, an Apart Research Sprint},
howpublished = {\url{https://apartresearch.com/sprints/projects/hikayat-interactive-stories-to-learn-ai-safety}},
url = {https://apartresearch.com/sprints/projects/hikayat-interactive-stories-to-learn-ai-safety}
}More from Women in AI Safety Hackathon
- Education track prizeView project: Morph: AI Safety Education Adaptable to (Almost) Anyone
Morph: AI Safety Education Adaptable to (Almost) Anyone
Morph
One-liner: Morph is the ultimate operation stack for AI safety education—combining dynamic localization, policy simulations, and ecosystem tools to turn abstract risks into actionable, culturally relevant solutions for …
- Mechanistic Interpretability PrizeView project: Red-teaming with Mech-Interpretability
Red-teaming with Mech-Interpretability
Red teaming large language models (LLMs) is crucial for identifying vulnerabilities before deployment, yet systematically creating effective adversarial prompts remains challenging. This project introduces a novel …
- Social Sciences track prizeView project: Detecting Malicious AI Agents Through Simulated Interactions
Detecting Malicious AI Agents Through Simulated Interactions
SafeAIGuard
This research investigates malicious AI Assistants’ manipulative traits and whether the behaviours of malicious AI Assistants can be detected when interacting with human-like simulated users in various decision-making …