Real Safety AI Foundation / Research

Children, Schools, and Youth Policy

Beyond Age Gates: A Privacy-Preserving Architecture for Predatory Pattern Detection in Youth-Facing AI Systems

Publisher: Real Safety AI Foundation

Working paper. Not peer reviewed.

Written
February 2026
Pages
9
Download PDFPlain language slides (PDF)Export BibTeXExport RIS

Abstract

Biometric age verification has become the default industry response to child safety concerns in digital communications, despite fundamental structural failures in legal compliance, technical resilience, and conceptual design. This paper presents an alternative architecture: an on-device, deterministic, rule-based behavioral pattern detection system that identifies predatory communication sequences by analyzing the sender's output patterns alone, without determining the recipient's age or identity, without collecting biometric data from any user, and without transmitting any data from the device to external infrastructure. The system uses finite-state machines, weighted rule graphs, and deterministic input normalization to detect documented predatory behavioral sequences including grooming cadence, isolation tactics, age-probing, boundary testing, and escalation patterns. A unidirectional cryptographically signed update mechanism enables rule evolution without creating any data egress channel. A comprehensive prior art survey across patent databases, academic literature, and commercial products confirms that no existing system combines all of these architectural properties. The architecture is presented as a platform-agnostic, embeddable software component positioned as a direct replacement for biometric age verification.

Keywords

  • child safety
  • predatory behavior detection
  • age verification alternative
  • on-device processing
  • deterministic pattern matching
  • privacy-preserving architecture
  • grooming detection
  • COPPA compliance

Plain language slides

First slide of the plain language summary of Beyond Age Gates: A Privacy-Preserving Architecture for Predatory Pattern Detection in Youth-Facing AI SystemsOpen the 21-slide summary (PDF)

Suggested citation

Gilly, Travis. "Beyond Age Gates: A Privacy-Preserving Architecture for Predatory Pattern Detection in Youth-Facing AI Systems." Real Safety AI Foundation Working Paper, February 2026. https://realsafetyai.org/research/beyond-age-gates/

References (11)

This list was read from the PDF text. Where the two differ, the PDF is correct.

  1. [1]International Business Machines Corp. (2012). Predator and abuse identification and prevention in a virtual environment. U.S. Patent 8,099,668 B2. Filed January 7, 2008.
  2. [2]Anti-Toxin Technologies, Inc. (2021). Detecting and identifying toxic and offensive social interactions in digital communications. U.S. Patent Application 2021/0234823 A1. Published July 29, 2021.
  3. [3]Street, J., & Olajide, F. (2023). Evaluating a non-platform-specific OCR/NLP system to detect online grooming. Proceedings of the 18th International Conference on Cyber Warfare and Security (ICCWS 2023), Towson University. Academic Conferences International Limited.
  4. [4]Federal Trade Commission. (2024). Children's Online Privacy Protection Rule; Final Rule. 16 CFR Part 312. Federal Register, 89 FR 2034.
  5. [5]Crown Prosecution Service. (2019). CPS strengthens guidance on child abusers caught in undercover stings. CPS Policy Guidance, United Kingdom.
  6. [6]U.S. Department of Justice. (Various). FBI undercover operations leading to federal charges for individuals attempting to meet minors for sex. DOJ Press Releases, Southern District of Illinois and others.
  7. [7]Microsoft/Thorn. (2020). Project Artemis: Grooming detection technology for chat platforms. Deployed via Thorn's Safer product for platform integration.
  8. [8]Vogt, M., Leser, U., & Akbik, A. (2021). Early detection of sexual predators in chats. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (ACL-IJCNLP 2021), Volume 1: Long Papers, pp. 4985-4999.
  9. [9]Bours, P., & Kulsrud, H. (2019). Detection of cyber grooming in online conversation. IEEE International Conference on Identity, Security and Behavior Analysis (ISBA 2019).
  10. [10]Chehbouni, K., et al. (2025). Enhancing privacy in the early detection of sexual predators through federated learning and differential privacy. arXiv preprint arXiv:2501.12537.
  11. [11]Technology Coalition. (2021). Online grooming: Considerations for detection, response, and prevention. Research Paper.

All research