Cite2Root
Dong Chen · Team Plswearpants
Submitted to AI Safety Entrepreneurship Hackathon. Sprint projects are early-stage work by participants, not Apart Research publications.
Regain information autonomy by bringing people closer to the source of truth.

Reviews
This approach currently doesn't seem to apply enough to core AI risk challenges, and it's unclear given the lack of empirical results given in this paper if things will scale or if this intervention actually works. The concrete tooling is appreciated though.
I enjoyed the thorough introduction to the problem space, and to recent innovations like Takehiko et .al When ti comes to the solution, the phases of production are clearly defined. I couldn't dive deeper into the MVP because the Github repo seems broken (https://github.com/Piswearpants/cite2root).
A proposal for the team is to also consider how to productize the solution, how to put in the market and design dynamics and incentives for its adoption.
Also ti reads as if the solution focuses on validating existing citations; what about identifying claims that require a citation (but don't have one yet)? Wikipedia worked on something similar: https://meta.wikimedia.org/wiki/Future_Audiences/Experiment:Citation_Needed
An issue that is in the market, but it's not clear how scalable the solution is. Go to market strategy isn't clear either. It is also unclear if they're able to identify hallucinations or just works as a "reverse" perplexity or a wrapper towards something that's able to solve some of the challenges around . This does not tackle AI safety risks as traditionally understood, but is definitely a pertinent area.
Cite this project
@misc{chen2025cite2root,
title = {{Cite2Root}},
author = {Dong Chen},
year = {2025},
month = jan,
note = {Submitted to AI Safety Entrepreneurship Hackathon, an Apart Research Sprint},
howpublished = {\url{https://apartresearch.com/sprints/projects/cite2root}},
url = {https://apartresearch.com/sprints/projects/cite2root}
}More from AI Safety Entrepreneurship Hackathon
- 1st place by peer reviewView project: AntiMidas: Building Commercially-Viable Agents for Alignment Dataset Generation
AntiMidas: Building Commercially-Viable Agents for Alignment Dataset Generation
the commonwealth
AI alignment lacks high-quality, real-world preference data needed to align agentic superintel- ligent systems. Our technical innovation builds on Pacchiardi et al. (2023)’s breakthrough in detecting AI deception …
- View project: Scoped LLM: Enhancing Adversarial Robustness and Security Through Targeted Model Scoping
Scoped LLM: Enhancing Adversarial Robustness and Security Through Targeted Model Scoping
FocusAI
Even with Reinforcement Learning from Human or AI Feedback (RLHF/RLAIF) to avoid harmful outputs, fine-tuned Large Language Models (LLMs) often present insufficient refusals due to adversarial attacks causing them to …
- View project: Prompt+question Shield
Prompt+question Shield
Seon's team
A protective layer using prompt injections and difficult questions to guard comment sections from AI-driven spam.