Digital Sentinel AI
Automated Regulatory Compliance Monitoring for Digital Food Marketing
Overview
Digital Sentinel AI is an independent doctoral research project — a Python-based regulatory compliance monitoring platform that automates the detection and assessment of digital food marketing content for signals of youth-targeting risk.
The platform addresses a critical gap in public health surveillance: despite rapidly evolving regulatory frameworks for digital food marketing, enforcement capacity has not kept pace. Systematic monitoring of brand compliance with youth-targeting restrictions at scale remains a largely manual, resource-intensive process. Digital Sentinel AI demonstrates that automated, evidence-based compliance monitoring is technically feasible at a fraction of the cost of manual review.
View Live PlatformThe Problem
Digital food marketing targeting children is a growing public health concern. Ultra-processed food brands — classified under NOVA Group 4 — increasingly reach children through social media, influencer content, and targeted advertising, often in ways that may conflict with emerging regulatory frameworks including:
- UK High in Fat, Salt and Sugar (HFSS) regulations — restricting the promotion of less healthy products in digital and broadcast media
- EU Digital Services Act (DSA) — establishing obligations around targeted advertising directed at minors
Despite these regulatory developments, public health researchers and regulatory bodies lack accessible tools to continuously track and assess digital marketing behavior across platforms at scale.
What the Platform Does
Digital Sentinel AI follows a four-stage automated pipeline:
Why It Matters
Regulatory frameworks for digital food marketing are evolving rapidly, but enforcement capacity has not kept pace. This platform demonstrates three things:
- For public health researchers — a replicable methodology for systematic digital marketing surveillance at scale
- For regulatory bodies and advocacy organizations — a demonstration that technology can support evidence gathering for policy enforcement
- For the broader field — a contribution to the emerging evidence base on the reach and nature of digital food marketing targeting children
Tech Stack
| Component | Technology |
|---|---|
| Web Scraping & Automation | Python, Playwright |
| Backend Framework | Flask |
| Evidence Notarization | SHA-256 Hashing |
| Deployment | Hugging Face Spaces |
| Data Processing | Python — pandas, NumPy |
| Monitoring Scope | NOVA Group 4 / HFSS brands |
| Regulatory Frameworks | UK HFSS, EU Digital Services Act |
Potential Applications
While Digital Sentinel AI was developed with a focus on NOVA Group 4 food marketing and youth-targeting risk, the underlying pipeline is adaptable to a range of public health surveillance contexts:
Regulatory Compliance Monitoring
Automated assessment of brand content against defined regulatory criteria — including HFSS, DSA, or equivalent national frameworks — to support enforcement bodies in identifying potential violations at scale.
Academic Research on Digital Marketing
Systematic, reproducible collection and classification of digital marketing content for peer-reviewed research on commercial determinants of health, including tobacco, alcohol, and ultra-processed food promotion.
Policy Advocacy and Evidence Generation
Generation of audit-ready, SHA-256 notarized evidence records that advocacy organizations and public health bodies can use to support policy arguments and regulatory submissions.
Multi-Country and Cross-Platform Surveillance
The modular pipeline architecture allows adaptation across platforms, languages, and product categories — making it applicable to multi-country research programs tracking digital marketing behavior across jurisdictions.
Longitudinal Marketing Trend Analysis
Continuous monitoring over time enables detection of shifts in marketing tactics, messaging strategies, and platform behavior — supporting both retrospective analysis and real-time public health intelligence.
Current Status
Digital Sentinel AI is live and deployed as an independent doctoral research project. The platform is actively monitoring digital marketing content and generating risk assessments across tracked brands.
Platform capability findings — including aggregate risk signal detection across monitored content — are currently being prepared for peer-reviewed publication. Specific brand-level findings will be reported through formal academic channels following peer review.
Limitations and Future Work
Current scope: - Monitoring is currently focused on a defined set of NOVA Group 4 brands - Classification thresholds are calibrated to HFSS and DSA criteria and may require adjustment for other regulatory contexts
Planned development: - Expand brand and platform coverage - Develop standardized reporting templates for regulatory submission - Validate risk scoring methodology through peer review - Explore applicability to tobacco and alcohol marketing compliance contexts