Skip to content
Apart Research
Apart Research
Contact us
Sprint projectNov 23, 2025Nairobi, Kenya

AI Sentinel

Nice Cailie Ineza · Team Middle

Submitted to Defensive Acceleration Hackathon. Sprint projects are early-stage work by participants, not Apart Research publications.

Read the report

Report: AI Sentinel

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

AI Sentinel is the first dual-domain monitoring system for AI outputs, detecting biosecurity and cybersecurity threats in real-time. Unlike existing tools that screen after synthesis requests or deployment, AI Sentinel intercepts threats at the AI generation layer with 120ms response time. Our three-layer architecture (bio + cyber + intent) achieves 88.2% accuracy on real-world benchmarks including CVE-2024 exploits and select agent detection. Key innovation: counter-intuitive jailbreak logic where "educational claims + dangerous content = HIGHER risk" catches bypass attempts that fool standard filters. Deployment scenarios include AI research labs (API middleware), DNA synthesis providers (upstream screening), enterprises (insider threat monitoring), and developer teams (IDE plugins). Open-source architecture enables community-driven improvements while complementing existing tools like SecureDNA and Nuclei.

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 *

Does this reduce AI-related catastrophic or existential risks?

Scoring guide
  1. 1Minimal Impact. The project has minimal relevance to AI safety. It doesn't meaningfully address risks uniquely enabled or accelerated by advanced AI systems.
  2. 2Tangential Connection. The project touches on AI safety concepts but lacks depth or specificity. The connection to AI-enabled threats (bio, cyber, or AI misuse) is weak or unclear.
  3. 3Clear AI Safety Value. The project clearly reduces AI-related risks with valuable contributions. It addresses specific threats from AI systems and engages meaningfully with biosecurity, cybersecurity, or AI safety challenges.
  4. 4Significant Impact Potential. The above, plus the project demonstrates scalable safety mechanisms or defensive approaches. It shows clear potential to buy time for solving harder problems like alignment, or creates positive externalities for the broader AI safety ecosystem.
  5. 5Major Advancement. The above, plus the project represents a significant leap forward in defensive AI safety. Judges would eagerly share this with biosecurity, cybersecurity, or AI safety researchers and expect it to influence the field.

Does this strengthen the shield against AI-enabled threats?

Scoring guide
  1. 1Minimal Relevance. The project is only tangentially related to defensive technology or societal protection. Connection to biosecurity, cybersecurity, or defensive infrastructure is unclear or missing.
  2. 2Some Relevance. The project has some relevance to defensive acceleration, but the connection is broad or generic. It touches on defense without specific focus on AI-enabled threats or protective capabilities.
  3. 3Clear Relevance. The project clearly addresses defensive gaps against AI-enabled threats. It connects to at least one track (biosecurity, cybersecurity, or defense infrastructure) and demonstrates understanding of the threat landscape.
  4. 4Strong Contribution. The above, plus the project builds on existing defensive approaches and offers novel tools, frameworks, or implementations. It explicitly explains how it strengthens defensive capabilities with realistic deployment potential.
  5. 5Breakthrough Impact. The above, plus the project provides breakthrough insights or tools that could significantly influence defensive technology development. It identifies critical gaps and presents compelling solutions with clear paths from prototype to deployed system.

Did you build something that actually works?

Scoring guide
  1. 1Incomplete or Flawed. The project appears rushed or incomplete. Technical implementation is flawed, core functionality doesn't work, or the approach is fundamentally unsound. Little to no documentation.
  2. 2Basic Competence. The project shows reasonable effort with basic technical competence. Core functionality partially works. Documentation exists but may be incomplete. Some limitations are acknowledged.
  3. 3Solid Hackathon Project. The project is technically solid and well-scoped for 48 hours. Core functionality works and is documented. Code/methods are understandable and limitations are honestly addressed. This is what a good weekend prototype should look like.
  4. 4Impressive Implementation. The above, plus the implementation exceeds typical hackathon quality. Clear methodology, thorough documentation, and working demo. The tool/prototype is immediately useful for defenders and could realistically be built upon.
  5. 5Exceptional Execution. The project far exceeds expectations with exceptional technical execution. The implementation is elegant, fully functional, and includes something special (e.g., deployed demo, exceptional documentation, innovative architecture, or clear startup potential).

  1. I like that it's a working tool for people to use for flagging, but the implementation for detection is fairly basic. Would have liked to see how this project adds to the existing

  2. Thanks for building this! The concept of using cheaper detection systems within a defence-in-depth framework is sound and the execution is impressive for a hackathon project.

    I think there are some significant limitations that cap its score. In particular:

    1. I agree that 'educational' context should *not reduce* risk scores, but I don't think it's obvious that it should increase scores. I think this might be a case of overfitting, where it works well in this case but not for the more sophisticated attackers from whom most of the harm could come (assuming that the distribution of AI harms is heavy-tailed).

    2. Similarly, the regex and phrase matching is very fast but limited. It might provide a false sense of security - sophisticated attackers could refer to the same concepts using different words to avoid the matching, given there isn't semantic generalisation etc.

    3. This implementation would appear to have a very high false positive rate in many real-world deployment settings, given that, as I understand it, e.g. 'sudo' and 'socket' trigger high risk scores.

    4. Bluntly, I think you're at risk of overselling what you have achieved - if you built this and it was adopted, the most obvious way it could be net negative is for customers is for them to think they're now safe, so not use other security measures. Although it's technically true that it's 'production-ready', has '88% accuracy' and 'catches [some] threats before they cause harm', this kind of language gives a greater impression of safety than can be achieved (I think) using these methods.

    5. The number of test cases is quite limited.

    But this is a great hackathon project and some of these issues are fixable, e.g. you could

    a. Experiment with more sophisticated filters that are still low-cost - what does the cost/accuracy frontier look like?

    b. 'Sell' this as a first defence that is very effective against a specific class of attacks but make clearer caveats about its limitations

    c. Do a wider range of tests

    Read full reviewShow less

Cite this project

@misc{ineza2025ai,
  title = {{AI Sentinel}},
  author = {Nice Cailie Ineza},
  year = {2025},
  month = nov,
  note = {Submitted to Defensive Acceleration Hackathon, an Apart Research Sprint},
  howpublished = {\url{https://apartresearch.com/sprints/projects/ai-sentinel-ounq}},
  url = {https://apartresearch.com/sprints/projects/ai-sentinel-ounq}
}

More from Defensive Acceleration Hackathon

All 84 entriesBrowse all projects

Build something like this at the next Sprint

AI Collusion Research Sprint · Oct 23 - 25, 2026